Steamturbines Church

Prof Alfred vi Ferrotti
UZlu.l^^Mi^
STEAM TURBINES
STEAM TURBINES
EDWIN F.''
/
CHURCH, Jr., S.B., S.M.
Professor of Mechanical Engineering, Polytechnic
Institute of Brooklyn
First Edition
McGRAW-HILL BOOK COMPANY, Inc.

NEW YORK: 370 SEVENTH AVENUE
LONDON: 6 & 8 BOUVERIE ST.. E. C. 4
1928
Copyright, 1928, by the
McGraw-Hill Book Company, Inc.
PRINTED IN THE UNITED STATES OF AMERICA
THE MAPLE PRESS COMPANY, YORK, PA.

PREFACE
This book was prepared because the author experienced diffi-
culty in finding a pubhshed text on Steam Turbines which was
satisfactory for a short but thorough course for juniors or seniors
in engineering colleges, following and applying the study of
thermodynamics. Of many excellent books on the subject,
the majority treat it too exhaustively for such class room use.
The subject of thermodynamics is so widely taught today that
was not thought necessary to include even a brief review. A
it
working knowledge of the ordinary steam cycles, the ability to

use steam tables and charts, and familiarity with the ordinary
laws of mechanics, are assumed. The opening chapter is
intended to acquaint the reader with the general form and
principal characteristics of a variety of turbines, and with the
meanings of the various terms to be used. Thereafter atten-
I p tion is centered on the application of thermodynamic principles
I to the calculation of the flow of steam through the nozzles and
Z blade passages of the turbine, the changes in its heat, pressure
'^^^""and volume, and the various factors affecting efficiency. Many
=»-; topics of interest and importance in turbine design have had
o< to be omitted for the sake of brevity.

Owing steam table values on the
to the changes probable in
"completion of the work of the A. S. M.
E. Committee on Steam
Research and Extension of the Steam Tables, no tables or charts
have been included, as is usual with turbine texts. Comments
,^;;en the various facilities now available are made in Chapter II.
'ô Illustrative problems have been given at the end of the text,
^classified under chapter headings. It has been the author's

['êxperience that classes progress most rapidly if problems are
Qpassigned sparingly during the early part of the text, extensive
use being made of the design problems when Chapter VII is
reached.
The author gratefully acknowledges his indebtedness to the
General Electric Company, the Westinghouse Electric and
V
vi PREFACE
Manufacturing Company, the De Laval Steam Turbine Com-

pany, the Allis-Chahners Manufacturing Company, the Terry
Steam Turbine Company and the Ljungstrom Steam Turbine
Company for supplying various data, drawings and photographs.
Many acknowledgements have been included in the text for
the use of equations and data, and it is hoped that none has been
overlooked.
Thanks are also due to Professor Frank D. Carvin of the
Polytechnic Institute of Brooklyn for his kind assistance in
reading the proofs.
E. F. C, Jr.
Brooklyn, N. Y.
December, 1927.
CONTENTS
Page
Preface v
Symbols ix
Chapter
I. Types and Characteristics 1
II. Flow of Steam in Nozzles 48
III. Blading 92
IV. Losses 170
V. Efficiency, Reheat Factor, and Condition Line 186
VI. Output Variation and Governing 205
VII. Design 219
Problems 260
Index 267
SYMBOLS
Most of the symbols to be used are assembled here for convenient refer-
ence. The meanings of subscripts for V are given in Sec. 33, and of sub-
scripts for E, e, h, Q, q and e in Sec. 72.
A = area, square feet or square inches.
A = 1 -^ 778 = reciprocal of mechanical equivalent of heat.
b = breadth of blading, inches.

c = clearance, inches,
c = coefficient of quantity discharge.
D = diameter, feet.
d = diameter, inches.
E = energy or work per pound of steam, B.t.u. or foot-pounds according to
context.
E = energy or work delivered by turbine as a whole, B.t.u. or foot-pounds
per pound of steam.
e = net heat drop or heat changed into work in an expansion, B.t.u. or
foot-pounds per pound of steam (see Sec. 72).
F = force or impulse, pounds.
uh = heat content of steam above 32°, B.t.u. per pound (subscripts i,q.,
sat., signify liquid and dry saturated states, respectively).
h = height or length of blades, inches.

U = intrinsic or internal energy, B.t.u. or foot-pounds per pound.
K = kinetic energy, foot-pounds per pound.
kn = velocity coefficient for flow through a nozzle.
kb = velocity coefficient for flow through blade passages.
m = thickness coefficient for blade or nozzle edges.
p = pound per square inch or per square foot.
Q = heat energy increment for a group of stages or for a turbine as a
whole, B.t.u. per pound (see Sec. 72).
q = heat energy increment for a single expansion or stage, B.t.u. per
pound (see Sec. 72).
R = reheat factor.
r = heat of vaporization, B.t.u. per pound.
r = pressure ratio p-y -^ p\.
To = pressure ratio po -^
P\.
S = arc occupied by nozzles, inches.
T = temperature, absolute degrees Fahrenheit.
t = temperature, degrees Fahrenheit.
V = velocity in feet per second (see Sec. 33).
V = specific volume, cubic feet per pound.
W = weight of flow, pounds per second.
X = percentage dryness of steam.
ix
X SYMBOLS
a = alplia = angle made by nozzle axis with direction of blade velocity.
/3 = beta = blade entrance angle.
7 = gamma = blade exit angle.
5 = delta = angle made by absolute exit velocity.
e = epsilon = efficiency (see Sec. 72).
K = kappa = ratio of specific heats = adiabatic exponent,
p = rho = ratio of blade speed to steam speed.

ip = phi = entropy.
= theta = per cent of circumference occupied by active nozzles.

T = tan = ratio of working steam to total steam.
Subscripts i and 2 signify initial and final states of an expansion.
Subscript signifies throat or critical condition in a nozzle.
Prime '
generally signifies conditions as modified by friction
STEAM TURBINES
CHAPTER I
TYPES AND CHARACTERISTICS
Turbines have developed to their present forms along

different lines according to steam conditions, type of service,
requirements as to cost, use, etc., and this process is still
going on, with all builders continually experimenting,
changing, and improving their product. There has been
no standardization of steam turbines comparable to that,
for example, in electric generators and motors, except to
a limited extent in the smaller sizes, and there is not likely
to be any such standardization in turbines unless and until
purchasers unite and standardize the conditions under
which power-generating systems will be run. Important
changes are even now taking place, with the advent of
exceptionally high pressures, stage feed-water heating,
reheating, development of extraordinarily large units,
and the trial on a large scale of the steam-mercury combina-
tion unit.
Before studying the theories underlying the design of a
steam turbine, it is obviously desirable to become
acquainted with at least a few of the many types that have
been evolved, so as to understand better the meanings of
the terms used. Some descriptions will therefore be given,
and as an aid in relating the various terms, and for ease in
later reference, they will be arranged in connection with a
series of classifications of turbines according to their char-
acteristic features.
1
:
2 STEAM TURBINES
1. Definitions.
A
steam turbine may be defined as a form of heat engine
in which the heat energy of the steam is transformed into
kinetic energy by means of flow through nozzles, and this
kinetic energy is in turn transformed into force doing work
on rings of blading mounted on a rotating part.
The usual turbine consists of four fundamental parts:
the rotor, which carries the blades or buckets; the casing,
cylinder, or stator, within which the rotor turns and in which
are fixed the nozzles or flow passageTfoi'THe^ steam and the ;
frame or base for supporting both the cylinder and the rotor,
the latter being carried in bearings. Cylinder and frame
are often combined. Other accessories necessary for suc-
cessful continuous commercial operation are a controlling
or governing system for adjusting the turbine output to
suit the load to be carried and for maintaining constant
speed, a lubricating system, piping for steam supply and
exhaust, and (for economical power generation in large
quantities) a condensing and evacuating system.
2. Classifications of Steam Turbines.

Steam turbines may be classified in the following ways:
^4. With respect to form of blade passage:

a. Impulse.
1. Simple, or single-stage.
2. Velocity stage.
3. Pressure stage, Rateau.
4. Combination pressure and velocity stage, Curtis.
b. Reaction, Parsons.
c. Combination impulse and reaction.
B. With respect to internal design and flow sequence:
a. Single-flow.
b. Single-double-flow or semi-double-flow.
c. Double-flow.
d. Divided-flow.
e. Compound, two- or three-cylinder, cross- or tandem-connected.
C. With respect to direction of steam flow relative to plane of rotation

a. Axial flow.
b. Radial flow.
C, Tangential flow,
: : :
TYPES AND CHARACTERISTICS 3
D. With respect to repetition of steam flow through bhides

a. Single-pass.
b. Re-entry or repeated flow.
E. With respect to rotational speed

a. For 60-cycle generators.
b. For 25-cycle generators.
c. For geared units, and for direct-connected or electric-drive marine
units, no especial speed requirements.
F. With respect to conditions of use

a. High-pressure condensing.
b. High-pressure non-condensing.
c. Mixed pressure.
d. Bleeder or regenerative.
e. Extraction.
/. Low-pressure or exhaust.
g. Reheating.
These will now be considered in detail.
3. Turbine Characteristics.
A. With Respect to Form of Blade Passages.

—
Impulse Turbine. The complete meaning of this term
a.
as applied to turbine blading will be considered later; for

the present it may be defined as a system in which all steam
expansion takes place in fixed nozzles and none occurs in
passages among moving blades.
A single-stage or simple-impulse turbine is illustrated in
Fig. 7^ and is shown diagrammatically in Fig. 1. Here the
steam expands from its initial to its final pressure in one
nozzle (or one set of nozzles all working at the same pres-
sure), resulting in a steam jet of high velocity which enters
the blade passages and, by exerting a force on them due to
being deflected in direction, turns the rotor. All velocity
and heat remaining in the steam after it leaves the single
row of blading is lost.
In the line diagrams representing the steam actions in
this and the following types (Figs. 1 to 6) different lines
represent the drop in pressure, the variation in velocity
and the increase in volume. As the descriptions are read
1 Trademark of the De Laval Steam Turbine Company.
4 STEAM TURBINES
and the diagrams compared, it will be noted that the steam

volume increases whenever the pressure decreases, but the
resulting velocity changes depend on the type of turbine.
As a matter of fact, these velocity changes are distinguishing
characteristics of the different types.
A velocity stage-impulse turbine (illustrated by Figs. 8,
10a and 106, and shown in diagram in Fig. 2) has one set of
nozzles, with several rows of blades following it. In pass-
ing from the nozzle exit through one set of blades the veloc-
ity of the steam is lowered by virtue of the work done on
the blades, but high.
is still It then passes through a row
of fixed guide blades which change the direction of the steam
until it flows approximately parallel to the original nozzle
direction, discharging it into a second row of blading fixed to
the same wheel. This second row again lowers the steam
velocity by virtue of the work delivered to the wheel. A
second set of guide blades and a third row of moving blades
are sometimes used.
In Fig. 8, the steam enters at the top through a pipe not
shown, through a steam strainer and governor valve into a
steam chest supplying the various nozzles spaced around a
portion of the periphery. Individual nozzles may be opened
or closed by a hand wheel (not shown) on the stem of
the nozzle-co7itrol valve. The turbine wheel is mounted on
a shaft which passes through the casing to bearings out-
side, carbon packing being used in the stuffing boxes of this
turbine to maintain steam tightness. The governor is
mounted on the right-hand end of the shaft and operates
the balanced piston governor valve through two rockers and a
connecting link. On the left end of the shaft is the coupling
machinery. In Fig. 106 are shown
for attaching the driven
and guide blades, which in this case
details of the nozzles
are attached permanently to the nozzles. In Fig. 135 is
shown an equivalent arrangement from a large turbine,
the nozzles in this case being contained in a curved nozzle
plate secured to the face of the nozzle chamber; the guide
blades are also attached to this chamber as shown. Figure
10a shows the rotor assembly, and the form and method of
attachment of the blading is well shown.
The inherent defect of the velocity-stage turbine (to
which class also belongs the re-entry turbine) is the exces-
sively high steanT velocity generated by a single expansion
from the initial to the back pressure, of the order of 3,500
ft.per second for expansion into a vacuum, requiring an
extremely high blade speed to utilize it efficiently and
resulting in a high rotor speed; the large steam velocity
also causes large frictional losses.
A pressure stage-impulse turbine is shown in Fig. 9, the
diagram being as shown in Fig. 3. Each of a series of
chambers, formed by parallel disc-shaped partitions,^ has
a simple-impulse turbine enclosed in it, all wheels being
fastened to the same shaft. Each chamber receives the
steam in turn through groups of nozzles placed on arcs,
the last chamber discharging to the condenser. The pres-
sure drop is divided into as many steps as there are chambers,
each being called a ''stage"; the resultant steam velocity

in each case is small, allowing reasonably low blade veloci-
ties and preventing excessive loss by steam friction. The
diagram (Fig. 3) shows that the steam pressure drops in
each stage and the steam volume increases; the steam
velocity is high at exit from the nozzles and is low at exit
from the blades. This arrangement is often termed a
Rateau turbine, and the separate stages, Rateau stages.
In Fig. 11, these pressure stages are shown in stages 2 to 5.
In Fig. 12 is shown a view of one of these pressure-stage
turbines with the top cover removed, making the wheel
assembly clearty visible. The group of wheels showing at
the extreme rear is a low-pressure turbine mounted on an
extension of the shaft. In Fig. 13 (section in Fig. 15) is
shown a similar view of another pressure-stage turbine in
which the shaft and lower halves of thediaphragms are
clearly visible.The upper halves of thediaphragms are
fastened within the top half of the casing and lift up with it.
^Called diaphragms.
6 l^TEAM TURBINES
Figure 18 shows a view of a pressure-stage turbine during

assembly and illustrates excellently the relation of the various
parts. Units of this type are made in the largest sizes.
Turbines using combinations of pressure and velocity
staging are widely used, and are illustrated in Figs. 19
and 21, Fig. 4 showing the diagram. The wheel in each
chamber carries two (sometimes three) rows of blading
instead of one. The turbine has as many pressure stages
as there are wheels, and each pressure stage has as many
velocity stages as there are rows of blades on the wheel in
that stage. This arrangement makes a smaller, shorter,
and cheaper turbine, at more or less sacrifice of efficiency.
Commercial turbines of this type are called Curtis turbines,
after the original inventor, and the individual pressure
stages, each with two or more velocity stages, are sometimes
called Curtis stages. The nomenclature in Fig. 19 should
be carefully read.
A widely used combination has numerous pressure stages,
but only the first wheel carrying two rows of blades.
Figure 22 shows the rotor of such a turbine, and Fig. 20 a
sectional drawing of a turbine with a similar rotor though
with a much smaller diameter for the two-row wheel.
Figures 13, 14, and 15 illustrate other turbines with the
same combination, and Fig. 11 shows in detail the nozzle,
diaphragm, and blade arrangements in such a turbine.
b. Reaction or Parsons Turbine. — In the turbines so far
described, the steam expands only in fixed nozzles, and flows
through passages between blades arranged in rows, giving
up its velocity to these blade rows and causing them to
rotate against resistance. In the widely used reaction
turbine proposed and built by Sir Charles Parsons, the steam
drops in pressure and expands while it is passing through
the moving blades as well as in its passage through the fixed
nozzles. Just how this is brought about and what is
gained by it will be studied presently, but it results in a
radically different form of turbine (as illustrated in Fig. 27
and in the diagram of Fig. 5). Blades are set in rows on
the surface of a rotor in the form of a cyhnder stepped in

several diameters (Fig. 28); the rows of blades on this
rotor re\'olve between rows of similar blades set on the inter-
nal cylindrical surface (also stepped) of the cylinder or
casing (Fig. 29). (The three figures are not of the same
turbine.) As the steam flows toward the left from the
inlet (Fig. 27), it expands, hence both the moving blades
on the rotor and the fixed blades in the casing increase in
height; at intervals it is necessary also to increase the
diameter in order to accommodate this increase of volume,
as the blade heights would otherwise become excessive.
The balancing pistons at the right of the rotor in Figs. 27 and
28 are for the purpose of counterbalancing the axial thrust
due to high steam pressure, as will be explained later.
The fixed and moving blades are of substantially the
same form (Fig. 148). Figure 158 shows the moving
blading in place in the rotor, and Figs. 159 and 160 show
the relative positions of moving and fixed blades. The
clearance between the tips of the blades and the rotor or
cylinder must be carefully kept at the minimum allowable
during the process of manufacture, and in operation the
rotor must be maintained accurately in its proper axial
position.
c. Combination impulse and reaction turbines, with a
velocity stage or two-row impulse wheel at the inlet end,
and reaction bladingfor the remainder of the unit, charac-
terize much product of the Westinghouse Electric
of the
and Manufacturing Company. The diagram is given in
Fig. 6. Figures 31 and 37 picture the rotors of two such
turbines, of moderate and large capacities, respectively;
corresponding sectional drawings are shown in Figs. 30
and having a special form of multiple-flow
35, the latter
blading (also calledBaumann blading, from the inventor)
at the low-pressure end, which will be explained in detail
later.
The advantage of this combination is that the compact
impulse wheel can replace the long and relatively slender
8 STEAM TURBINES
high-pressure drum of the straight reaction turbine (com-

pare Figs. 28 and 31), decreasing length, space occupied,
cost, and tendency to vibrate and making it more satis-
factory in many ways, though also decreasing its efficiency.
Certain advantages are also gained in connection with
governing (Chap. VI). The impulse wheel is, therefore,
much used in this combination, though in some of the very
largest turbines the reaction principle is retained in the
high-pressure stage also (Fig. 42).
B. With Respect to Internal Design and Flow Sequence.
a. In single-flow turbines the steam enters at one end and
flows once through the blading in a direction parallel to the
axis, emerging at the other end of the turbine and entering
the condenser. This is the most common arrangement in
small and moderately large turbines. Figures 7 to 11,
16, and 19 to 22 are good illustrations of this arrangement.
b. The single-double-flow turbine is used with reaction
blading in cases where it is desired to balance partially the

end thrust on the rotor, or where the quantity of steam is
too great to be passed through one set of low-pressure
blading of reasonable height. After entering near one
end and passing through the high-pressure and intermediate
pressure blading, the steam is divided into two equal por-
tions; one part continues on in the same direction through
a set of low-pressure blading, and the other part returns
through a hollow rotor to the end at which the start was
made and then passes through another set of low-pressure
blading (Figs. 38 and 40). The return may also be made
through passages formed in the casing (Fig. 42, left-hand
or low-pressure unit) or through an external cross-over pipe
(Fig. 41).
c. In double-flow turbines, which are generally low-pres-
sure units, the steam enters at the center and divides, the
two portions passing axially away from each other through
separate sets of blading on the same rotor. This type of
unit is completely balanced against end thrust and gives
large area for flow through the two sets of low-pressure

blading. Figure 32 shows a view of a double-flow rotor,

and Fig. 44shows a drawing of a double-flow unit having
Baumann blading.
In both the double-flow and the single-double-flow types,
the two low-pressure exhausts connect to the condenser at
separate points (Fig. 38), or are brought together to a single
connection by a Y-casting.
d. have been used in marine installa-
Divided-floiv units
tions for supplying steam to two turbines on separate
propeller shafts. After passing through high-pressure
and intermediate-pressure blading in a turbine on one shaft,
the greater part of the steam is sent to a low-pressure
turbine on another shaft, passing through it to the condenser.
The small remainder passes through a set of low-pressure
blading on the first shaft and into the same condenser.
The parts are so proportioned that approximately equal
powers will be generated on the two shafts.
e. Compound turbine is the term applied to turbines in
which the steam passes in sequence through two or more

separate units, expanding in each. Usually the high-
pressure and the intermediate-pressure portions are in one
casing, and the low-pressure portion in another. Two
units may be arranged tandem compound, in axial line on
the same shaft and bed plate and driving the same electrical
unit, or cross-compound, on separate shafts and bed plates
and driving separate electrical units. In the latter case
the low-pressure unit may rotate at a slower speed, which
is advantageous, but the speeds must have a certain definite
relation dictated by frequency requirements. When run-

ning, the two units are operated as one, being held rigidly in
step by the electrical phase relation. Figures 12 and 42
show tandem arrangements, and Figs. 43, 44, and 46, cross-
compound arrangements. In Fig. 45, there are two low-
pressure units. Three units in series may also be used, as
in Fig. 47, where the high-pressure and intermediate-pres-
sure units are arranged tandem, and the low-pressure unit
is separate.
10 STEAM TURBINES
The two turbines of a cross-compound unit may also be
geared to a single electrical generator (Fig. 46). In some
marine turbines, separate high-pressure, intermediate-
pressure and low-pressure units have been gear-connected
to a single propeller shaft.
C. With Respect to Direction of Stearn Flow Relative to
Plane of Rotation.
a. Axial flow signifies steam flow substantially parallel
to the axis of rotation among blades which are set radially.

This is the only arrangement used in medium and large
turbines, and is generally used in small turbines also.
It provides opportunity for almost any desired degree of
expansion of the steam by increase in the number of rows
and the length of the blades. All of the Figs. 7 to 47,
except 25 and 26, are axial-flow turbines.
Attention may here be called to the fact that in no turbine
does the steam whirl around with the rotating part, follow-
ing its direction. It is also of interest to know that the
length of time required for a given particle of steam to
pass through the turbine is exceedingly brief, being of the
order of magnitude of a few thousandths of a second.
b. Radial flow is obtained when the steam enters at or
near the shaft and flows substantially radially outward

among blades which are placed parallel to the axis of
rotation. In one turbine of this type, illustrated in principle
in Figs. 49 and 50, alternate rings of blades are attached to two
discs mounted on separate rotors turning in opposite direc-
tions and driving separate generators. Figure 48 shows a
section of the Ljungstrom turbine. Fig. 51 a transverse section
through the blading and Fig. 53 a view of the turbine and
the inner or commutator ends of the two generators driven
by the turbine rotors. In the detail Fig. 55, the rows of
blades are made up in strengthening rings (3) like
(1)
''squirrel cages" and are alternately attached to the right
and left discs (8) by expansion rings (5) of dumb-bell-shaped
cross-section. Leakage from row to row is prevented by
the tightening strips (4). In Fig. 48 these blade rings (42)
are shown assembled and connected to the inner and outer

parts of the discs (35) and (36) carried by the hollow hub
(40) on the turbine shaft (26) and (27). The back of the
turbine discs (35) and (36) carries a disc (34) having on its
face labyrinth packing engaging with corresponding pack-
ing on the fixed disc (33) ; this reduces steam leakage out-
ward from (40) and is made use of to counterbalance the
thrust. The packing is shown in detail in Fig. 56.
Leakage along the shaft is prevented by packing (25),
shown in detail in Fig. 54.
Steam enters through the branched supply pipe (17)
and through the openings (23) and (40) to the inside of the
innermost blade ring, then passes radially outward over the
whole 360 deg. through the blade rings (42) and the divided
blade passages (43) to the condenser, to which the casing
(19) connects. At first the increase of flow area naturally
provided by the increasing diameter of the blade rings is
too great and the blade lengths have to be diminished then ;
the steam commences to expand at a greater rate, requiring

longer blades. Figure 52a shows that, for condensing
turbines, the blades tend to become so long that they would
deflect, and in addition the direction of the steam flow
would no longer be even approximately radial. Figure 526
shows the method of subdivision which obviates these
difficulties. In turbines of large power, the maximum
possible flow area obtained by this method is still insuffi-
cient, so that radial blades are necessary. These are

arranged for double flow, and are illustrated in Figs. 57
and 58.
There are no fixed nozzles, so that expansion takes place
in both sets of blades, and the turbine is almost necessarily
of the reaction type. Owing to the high relative speed of
the alternate rows of blades, due to their opposite rotation,
only a moderate number of rows is required for complete
expansion with high efficiency. The elaborate packings
required to prevent leakage, the high cost of construction,
and the limitation on the size which can be built, however,
12 STEAM TURBINES
have all hindered development and few of these machines
are in use in this country.
c. Tangential flow is the term applied when the steam
enters through a nozzle placed approximately tangent to
the periphery and directed into semicircular buckets milled
obliquely into the edge of the wheel. Coupled with this is
the action of a reversing chamber fitted closely to the periph-
ery of the wheel, which has similar buckets milled into
it (Figs. 25 and 26) . These latter buckets receive the steam
discharged from the wheel buckets and return it again a
number wheel buckets in the proper direc-
of times to the
tion to produce additional work, the steam following an
approximately helical path. Any number of nozzles may
be used, each with its reversing chamber, up to that num-
ber which will completely fill the periphery,
D. With Respect to Repetition of Steam Floiv through
Blades.
a. Single Pass and (b) Re-entry or Repeated Flow. — Practi-
cally all turbines are single pass, except the tangential-flow
type just described, and the re-entry turbine, illustrated in
Figs. 23 and 24. In Fig. 24 the blades are set radially and
the steam, after passing once through them, following its
discharge from the nozzle, is returned by a reversing cham-
ber and again sent through them. Figure 23 shows such a
turbine wheel driving a generator by single reduction gear-
ing; the nozzle and reversing chamber appear at the lower
right-hand portion of the sectional view.
Re-entry turbines also fall into the velocity stage
classification.
E. With Respect to Rotational Speed,
a. QO-cycle Generators and (b) 25-cycle Generators. —Rota-
tional speed requirements are extremely rigid in turbines
which are direct-connected to electric generators, as these
must be a.c. units, sizes, and must,
except in the smallest
therefore, run at speeds corresponding to the standard
numbers of cycles per second. A 2-pole unit for 60-cycle
service runs at 3,600 r.p.m. and this is the maximum possi-
ble speed, as no generator can have than two poles. A

less
4-pole machine runs at 1,800, a 6-pole at 1,200 and an 8-pole
at 900 r.p.m. For 25-cycle current, a 2-pole machine runs
at 1,500 and a 4-pole at 750 r.p.m.
(c) Geared units may have practically any speed ratio
between the turbine and the driven machine, so that the

turbine may be designed for its own most efficient speed.
Many small-capacity geared outfits are used to drive small
generators, fans, pumps, etc., and a few large ones are used
for electric generators, but practically the only field for the
large geared unit has been in ship propulsion, where high
propeller efficiency requires slow or moderate propeller
speed (90 r.p.m. for slow speed ships up to 250 r.p.m.
for those of high speed.) Single-reduction gearing, con-
sisting of a pinion on the turbine shaft running with a large
gear on the propeller shaft, is used for moderate speed
reduction, say 7 or 8 to 1. In double-reduction gearing
another pinion and gear are added, allowing a speed reduc-
tion of 40 or 50 to 1. Figures 23 and 46 illustrate single-
geared arrangements.
F. With Respect to Conditions of Use.
a. High-pressure condensing turbines, built in the largest
sizes for power-plant use, and of the highest efficiency, are
supplied with steam of high pressure and temperature and
small volume, and exhaust it at very low pressure and large
volume. This great increase in steam volume calls for

short blading at the high-pressure end, and very high blading
at the low-pressure end (Figs. 9, 12 to 17, 27 to 30, 35 to 47).
b. High-pressure non-condensing turbines are used where
water for condensing purposes is unavailable, or where all

the exhaust steam at atmospheric pressure can be used for
factory processing or for heating. The steam volume at
exhaust moderate. These turbines are built in small or
is
moderate sizes only (Figs. 8, 19, 21, 23, 26).

c. Mixed-pressure turbines are supplied with steam
not only from a high-pressure source, but also with low-
pressure or exhaust steam, generally from another engine
14 STEAM TURBINES
or turbine. The low-pressure steam enters and joins with

the high-pressure steam after the hitter has passed a number
of rows of blades and has expanded down to the pressure at
which the low-pressure steam is ordinarily received.
Generally the low-pressure steam is expected to furnish
most of the energy, the controlling system supplying steam
from the high-pressure source to the high-pressure end only
when the load requirements are greater than can be met
by the low-pressure steam alone,
d. Bleeder or regenerative turbines, practically always
condensing, are in a sense the reverse of the above; all of
the steam is received from a high-pressure source, but at
one or more points before the steam has expanded to the
lowest pressure, steam is withdrawn at intermediate pres-
sures. Thus there are provided several supplies of steam
at different temperatures which are used for heating feed
water by stages, the feed water going first through a heater
supplied by the lowest temperature bled steam, then
through one with higher temperature steam, etc. This is
termed regenerative feed heating (Figs. 13, 15, 16, 35, 42, 43).
e. Extraction turbines, very similar to bleeder turbines,
are utilized in manufacturing plants where steam at about

atmospheric (or any other desired) pressure is required for
factory heating or processing. A turbine of this type is
characterized by a diaphragm or partition across the path
of the steam just beyond the intermediate point where the
bleeder connection is made, with ports through the dia-
phragm opened and closed by a governor-controlled valve;

sometimes an outside connecting pipe with control valve
replaces the ports in the diaphragm. When the ports (or
the control valve) are partially closed, the pressure in front
of the diaphragm rises and more steam is discharged into the
bleeder connection. If no bled steam is required, the gover-
nor opens the valve wide and the turbine operates as a

straight high-pressure condensing unit (Figs. 33, 34).
/. Low-pressure or exhaust turbines have come into use in
several ways. First, when very large powers are required

from a single machine, it is sometimes advisable to divide

the total expansion of the steam into two stages in separate
casings, resulting in a high-pressure and a low-pressure
unit (Figs. 12, 42, 43 and 44). In some cases two low-
pressure units divide between them the steam from one
high-pressure turbine (Fig. 45). In Fig. 47 is shown a
case where the total expansion is carried out in three units,
a high, an intermediate and a low pressure, the latter being
on a separate shaft. Second, in the stage of transition from
reciprocating engines to turbines, low-pressure units were
installed to receive the exhaust from reciprocating engines
at about atmospheric pressure and, through their ability
to utilize high vacua, were able to generate a large amount
of additional power at high economy. Third, certain
power requirements, as in steel rolling mills, are best served
by a reversing unit, and as turbines cannot be reversed, a
reversing reciprocatingengine is used, operating non-
condensing. This periodically furnishes large quantities
of atmospheric exhaust steam which, passed through a
heat-storage device called an accumulator, can efficiently
operate a low-pressure or exhaust turbine.
Low-pressure turbines are generally, though not neces-
sarily, made double flow.
g. Reheating turbines have been provided to meet the
demand for more economical power-plant operation. High

pressure steam is expanded in a turbine to an intermediate
pressure, then is withdrawn, sent back to the boiler room
for reheating (resuperheating) thus increasing its heat
content at the lower pressure, and is returned to the turbine
to complete its expansion. This results in a turbine
divided into two parts. Figure 17 shows an example
where the two parts are in one casing, and Fig. 47 an exam-
ple where three units, high-pressure, intermediate-pressure,
and low-pressure, are used, the steam being reheated
between the high-pressure and the intermediate-pressure
units. Provision may be made for a second or a third
reheating but at the lower pressure sthe steam volume is so ^
^Text continued on page 44.

16 STEAM TURBINES
/>A.
'S ni ni
Â- ri^ ?î2

/'/-es
K—^' A ye/
'
xî^^y î
Fig. 1. — Simple Fig. 2. — Velocity stage. Fig. 3. — Pressure stage,

impulse. (Curtis.) {Ratâu.)
\yoi
Fig. 4. — Pressure and velocity stage. Fig. 5. — Reaction. (Parsons.)

(Curtis.)
Fig. 6. — Combination velocity stage and reaction. Fig. 7. — Simple impulse

turbine.
ST£AM STRAINER GOVERNOR VALVE

BLADES
EMERGENCY NOZZLE
STOP VALVE NOZZLE CONTROL
- VALVE
Fig. 8.— Velocity-Stage turbine. (Dc Laxal Steam Turbine Company.)
Fig. 9. — Pressure-stage impulse turbine. (De Laval Steam Turbine Comjmny.)

18 STEAM TURBINES
5EC0ND ROW FIRST ROW OF BLADES
*, GOVERNOR WEIGHTS SPRIHG MD PIN
Fig. lOa.^ — Two-velocity-stage rotor. (De Laval Steam Turbine Company.)
HOZZLES WITH ATTACHED GUIDE BLADES
NOZZLE CONTROL VALVE
STEAM SUPPLY
Fig. 106. — Nozzle parts. {Dc Laval Steam Turbine Cotnpatiy.)

20 STEAM TURBINES
Fig. 12.Â50,000-kw., 20-stage l.SOO-r.p m taiulcin coinpound-iniiixilse turbine

with double flow at low-pressure eiul. ((Ji lu nd Eliclric Coinpany.)
Fig. 14. —A 20,000-kw., 13-stage, 1,500-r.p.m. impulse turbine. {General

Electric Coinpany.)
Fig. 15. — A 13-stage hloodor turhine, ty]>o R. {General Electric Company.)

22 STEAM TURBINES
Fig. 17. —A 35,000-kw. 1,800-r.p.m., 19-stage reheating-type impulse turbir

{General Electric Company.)
Fig. is. — Rotor, casing, and diaphragms of 60,000-kw. turbine. {General

Electric Company.)
24 STEAM TURBINES
uoraSEna Sn.€
ii connecfion Piece
30 Bearing BraCwe' tLPEna)
3 2 PaCfwng Tor DiQpn'-t

5 3 wneei' Casing (LOwer i
3<l Support iHPEnal
Fig. 19. — Curtis turbine, three pressure stages, each with two velocity stages.
{General Electric Company.)
Fig. 20. —A 5,000-kw., 5-stage, 3,600 r.p.m. Curtis turbine. (General Electric
Company.)
TYPES AXD CHARACTERISTICS 25
—
Fig. 21.- Curtis turbine, three pressure stages. Cover lifted, showing shaft,
bearings, packings, wheels, diaphragms, and guide blades. {General Electric
Company.)
Fig. 22. —A 500-kw., .j-stago Curtis turbine. Wheels on shaft. {General

Electric Company.)
STEAM TURBINES
I
2§
28 STEAM TURBINES
BAL^'NCINO PISTONS
Fig. 27. — High-pressure, single-flow reaction turbine. (Wctitinghounc Electric

and Maiiiifacturino Company.)
Fig. 28. —A 500-hp., condensing reaction turbine, rotor. {Aliis-Chahnm

Mannfarturina Company.)
Fig. 29. —A 7,500-kw. 1,500-r.p.m. reaction turbine, top half of cylinder with
blading installed. {Allis-Chabners Manufacturing Company.)
—
30 STEAM TURBINES
Fig. 31. C'oiiiMiiation impulse and reaction turbine, rotor. (Westinghouse

Elcdric and Manufaduriin} Company.) (See also Fig. 30.)
Fig. 32. — Rotor of low-pressure unit of cross-compound reaction turbine.

(Westinghouse Electric and Manufacturing Company.)
Fig. —
33.^ A 2,000- to 3,000-kw. combination impulse and reaction turbine,
bleeder type. {Westinghouse Electric and Manufacturina Company.)
Fig. 34a, b and c. — Bleeder valve and butterfly valve for turbine of Fig. 33.
32 STEAM TURBINES
TYPES AND CHARACTl'JIUSTICS 33
Fig. 36. —View of lower casing from low-pressure end, spindle removed.
Baumann type guide blading. {Westinghouse Electric and Manufacturing
Company.)
Fig. .37.^Rotor for turbine in Fig. 3.5, showing dummy pistons, high-pressure
impulse stage and reaction blading. {Westinghouse Electric and Manufacturing
Company.)
34 STEAM TURBINES
Fig. 38. — Single double-flow reaction turbine with high-pressure impulse

stage. Hollow spindle. (Westinghouse Electric and Manufacturing Company.)
(See also Fig. 40.)
Fig. 39. —A 6,000-kw., 3,600-r.p.m. combination impulse Company.)

and reaction turbine.
(Westinghouse Electric and Manufacturing
Fig. 40. — Single douhle-flow reaction turbine with high i)rossure impulse
stage. Hollow spindle. {Westinohouse Electric and Mantifactariny Company.)
(See also Fig. 38.)
Fig. 41. — Single douhle-flow reaction turl)inc, external cross-

inghouse Electric and Manufacturi/uj (Joinpanij.)
36 STEAM TURBINES
*tOffif3A09
1 r
I
Fig. 43. — High-pres.sure unit of cross-compound reaction turbine, with
impulse wheel. (Westiuohouse Electric and Maiuifacturinq Company.)
Fig. 44. —Low-pressure unit of cross-compound reaction turbine, double flow,

Baumann blading. {We sting house Electric and Manufacturing Company.)
38 STEAM TURBINES
L.P. Turb/na
@) ;
/f P. Turbine
Sf earn Supply
L.P Turbine
Fig. 45. —A 70,000-kw., cross-compound turbine. Interurban Rapid Transit

("ompany, New York. {Wcxtinghoiisc Electric and Manufacturing Company )
Fig. 46.^—A 4,000-kw., cross compound geared turbine driving direct current
generator. (Dc Laval Steam Turbine Com.pany.)
„ ,w
auijjooij ;u3nhsi>g f~^~'J |
1 ^
S ?
S 5
o s
JT 2
40 STEAM TURBINES
Fig. 48. — Section of 2,000-kw., Ljungstrom radial-flow turbine.

Fig. 51. — Section through blading system of Ljungstrom turbine. (Rows reduced
in number for sake of clearness.)
42 STEAM TURBINES
LJUNGSTROM TURBINE DETAILS
Fig. 52. —
Excessive length of rio. 5o. — Mew of turbine and generators,
low-pressure blade passages.
Fig. 54. — Shaft packing. Fig. 56.^ — Packing between fixed and rotat-
ing labj'rinth discs.
Fig. 57. — Section through upper part of a 14,000-kw. Ljungstrom turbine with
radial low-pressure blading.
Fig. 58. — View of Ljungstrom turljiae with radial low-pressure blades, cover
and guide blades removed.
44 STEAM TURBINES
large that the piping and control valves become large and
expensive, and the economical advantage doubtful.
The preceding descriptions are sufficient evidence of the
variety of turbines produced to meet the variety of require-
ments of users. At least one firm has put into operation a
system of manufacturing turbines, illustrated by Figs. 13
and 15, which will meet the large variety of demands
over a certain rather wide range, without requiring that a
separate special unit be designed and built to satisfy each
set of conditions. It will be noted that the casing of this
turbine consists of three parts fastened together by bolted
head containing the high-pressure packing, steam
flanges; a
admission arrangements, and first three stages of expansion;
a central portion or barrel containing five stages of expan-
sion ;
and an exhaust end containing five stages of expansion,
the low-pressure packing and the exhaust connection. The
company has prepared designs for five heads, three barrels,
and five exhaust ends, of varying diameters, lengths, and
numbers of stages; the large diaphragms allow variation
in wheel size, and by omitting a stage at any point a bleed
connection and control valve may be added. This system
provides for 125 combinations, varying from 1,500 to
10,000 kw., from 100 to 400 lb. initial pressure, and from 29-
in. vacuum to 70-lb. back pressure.
4. Advantages and Disadvantages of Steam Turbines.
The advantages of turbines over reciprocating engines are

very great. Less space is occupied, requiring smaller foun-
dations and building, the original cost is less, the mainte-
nance cost is less, the economy is greater (in part due to the
ease with which high vacua and high superheats may be
utilized), and they may be built in units of much greater
size, decreasing the cost of attendance. Engineers engaged
in designing small plants often find that it is there prefer-
able to use reciprocating engines, on account of the better
economy of these small sizes and the usual necessity for
reduction gearing in small, high-speed turbines, but the
reverse is usually true, and reciprocating engines of all
sizes have been scrapped and replaced by turbines long

before their useful lives were over. When it is admitted
therefore, that certain difficulties still meet the designer,
the builder, and the operator of steam turbines, it is not
to be understood that the overwhelming superiority of the
turbine to the reciprocating engine for most power-plant
purposes is in any way questioned.
Not all of the advantages the turbine possesses have been
secured at once, and the improvements in steam-turbine
design and construction, rapid as they have been, have
required the development of new methods in forge and
foundry practice, the design and construction of massive
shop equipment capableof finishing parts^t~o1j, degree of
accuracy formerly thought entirely impracticable, the
production of a multitude of metals having certain very
special properties, and the accumulation of a vast amount of
information drawn from practical experience, experimental
investigation, and scientific research. \ Experience in opera-
tion has alsobeenaiiimportai^^ and theie has been
close cooperation be^^^^^liej^'ê builders and the large
userôfêaJA-Jtu^Silies,^ The result is, that the nianu-
facture of turbines, except in^îhe^îâllestsi^Si^ is in the
hands'of a few large maniifacturers ^vjo, ha>Je beenr able to
survive the expense of this development work, and who, as a
result, have in their possession an enormous amount of
turbine data.
The operator is usually as much interested in reliability
as he is economy, and the removals of turbines from
in
service, or ''outages" as they are called, are due to ''rubs,"
or their effects, more than to any other cause lying in the
turbines themselves. The speed of wheel rims relative to
the adjacent stationary parts varies from 400 to 1,000 or
more feet per second; if contact does occur, so much heat
is generated almost instantaneously that the deflection
which caused the initial rub is apt to be increased, some-
times resulting in the complete destruction of the unit.
These rubs may result from one of several causes, even in a
46 STEAM TURBINES
turbine which is properly designed in the first place: (a)

improper adjustment of the rotor in its axial position, (6)
unbalance of the rotor producing excessive vibration and
deflection, (c) sudden wear of the thrust bearing which
holds the rotor in axial location, or of the bearings support-
ing the rotor, caused by failure of the lubricating system,
and (d) sudden changes in temperature produced by sud-
den changes in load conditions, causing unequal expansion.
In improperly designed or constructed turbines, the rela-
tive expansion of the rotorand casing may be so much that
with the clearance originally provided, it is impossible to
have any adjustment that will surely prevent rubbing at all
times. Also the discs or diaphragms may be mechanically
weak and may deflect under centrifugal force or steam
pressure, respectively. The wheels particularly will also
vibrate under certain conditions which must be avoided.
Corrosion and erosion of the blades and nozzles will usu-
ally necessitate their replacement after a period of service
whose length depends on the conditions of operation. In
some cases this has not happened within the commercial
life of the machine, i.e., the machine has become obsolete.
The economy of units seems to hold up fairly well through

years of service, corrosion and erosion causing a small
reduction with age, as does the increased leakage due to
increased clearance.
Mechanical losses are small, hence, the mechanical effi-
ciency ishigh and while perfection of the lubricating system
is practically essential for successful turbine operation,
when once it is and adjustments
attained, bearing troubles
are infrequent. Care of the unit resolves itself largely
into careful watching of the oil temperatures and proper
care in cleaning and renewing the oil.
Turbine repairing usually cannot be done by the operat-
ing staff in the plant, as it requires a considerable number
of extra men
with special technical training; hence, service
organizations for this purpose are maintained by large
builders. Extensive repairs require the return of the rotor,
the cylinder, or of the complete machine, to the builder's

plant, and as these, particularly the rotor, are massive and
heavy and at the same time have delicate parts and are
in size often close to the limits allowed by railroad trans-
portation, repairs are costly.
:
CHAPTER II
FLOW OF STEAM IN NOZZLES
From the preceding illustrations of actual turbines, it

will be evident that the primary purpose of a steam nozzle
in a turbine is to transform the greatest possible portion of
the available heat energy of the steam flowing through it
into the kinetic energy evidenced by the high velocity of
the issuing jet. By available heat energy is meant that
portion of the total contained heat energy which a perfect
engine, when operating between the temperatures at the
inlet and at the discharge of the nozzle, could change into
work. In addition, the nozzle performs the important
function of directing this jet on the moving blades in such
a manner as to allow an efficient utilization of this kinetic
energy by the blading.
6. Nature of Action in an Ideal Nozzle.

If the nozzle is to perform the above transformation with
perfect efficiency, it is evident that there must be no fric-
tional loss due to rubbing of the steam stream along the

nozzle walls, and there must be no losses due to eddy cur-
rents. The first requirement means walls of perfect
smoothness; the second, nozzle proportions and curvature
such that there is perfect stream-line flow. Furthermore,
there must be no conduction of heat through the nozzle
walls, to be lost to external surroundings by radiation, and
no conduction of heat along the nozzle walls to be
given back to the steam irreversibly at a lower pressure and
temperature. Under these ideal conditions, the action
would evidently be adiabatic; it may also be shown to be
reversible, as follows
48
FLOW OF STEAM IN NOZZLES 49
Assume a chamber A with a nozzle from which steam is
issuing with the frictionless adiabatic flow just described.

Imagine an exactly similar nozzle and chamber reversed in
mouth brought close to that of ^.
position, with its nozzle
The jet issuing from A contains, by virtue of its kinetic
energy, allmade available by the drop of pressure
the heat
from Pi to passing into the nozzle of B and coming
P-y. On
to rest in the succeeding chamber, this kinetic energy will
disappear, reappearing as an equivalent amount of heat
energy. By properly con-
trolling the discharge from
B, the pressure in can B
be made to rise to that in
A, and the steam will be
in its original state.
Under these ideal con-
ditions, the action is both
adiabatic and reversible, and the entropy therefore remains
constant.
No actual steam nozzle fulfils these ideal conditions,
but advantageous to derive the fundamental relations
it is
on these assumptions, and later to introduce such correc-

tions or modifications as may be necessary to make them
apply to actual working conditions.
In accordance with custom the term 'âdiabatic"
applied to flow through a nozzle will usually mean reversible
adiabatic flow.
7. Fundamental Energy Relation.

In order to derive the fundamental equation relating the
change of state of the steam in passing through the nozzle
to the kinetic energy developed at its exit, suppose that
the flow taking place from the larger pipe A into the
is
smaller pipe B (Fig. 62) that pi, Xi (or ti, if superheated)

;
are the conditions of the steam in A, and p2, x^ those of

the steam in B. If frictionless pistons be imagined in each
of the cylinders, it is evident that the piston in A will exert

'
50 STEAM TURBINES
a specific pressure pi on the steam in front of it, and each
pound of steam passing from A may be said to have the
work El = piVi done on it by this piston. Similarly, each
pound entering the cylinder B will do the work Eo = P2V2
on the smaller piston, since the steam in front of it exerts
a pressure p^ on it. Also, if Vi and V2 be the velocities
of the steam in A and B,
respectively, the corre-
sponding kinetic energies
Ki and K2 will be Vl/2g
Fi«- 62-
and Vl/2g.
Summing up the energies on the two sides of C and equat-
ing their totals, we will have
Internal energy 1 + external worki -\- Ki = internal

energy2 + external work2 + K2, (1)
-
Ijif-^ti^'Jj
which for steam becomes
^
^'^^
til + Apivi + K, =d2 + AP2V2 + K2, (2)
or
which is equivalent to
_
(VI V\\
= hi-h2 = q^.,ÛrL^
M^-iv")
V2sr 2g)
(4)
where qa is called the "availableheat" and hi and /i2 are

the heat contents per pound
steam at the beginning
of
and the end, respectively, of the reversible adiabatic expan-
sion. The term ''heat content" or h, as used in this text,
includes the feed-pump work, and has the same significance
as the expressions ''thermal potential" and "enthalpy"
used by some authorities, that is,
h = ic+ Apv.
8. Representation of Action on Heat Charts.
Since, as previously stated, the ideal expansion 1 to 2 is
isentropic, the value of hi — h^, which is here derived for the

case of flow through a nozzle, is identical with the heat equiv-
alent of the work done by a steam engine operating on the

Rankine cycle between the pressures pi and p2. In the case
of the reciprocating engine, this work is performed directly
on the piston and, transmitted by the connecting rod and
crank, appears as work done in rotating the shaft against
external resistance. In the turbine, the same work is
performed by the steam in accelerating the flowing steam
itself;^ as this swiftly moving steam passes through the
moving blades, its velocity and, therefore, its momentum,

decreases, the corresponding energy decrease reappearing
as work done in rotating the shaft against external resist-
ances. Many of the conclusions and relations already
derived (in thermodynamics) for the case of the Rankine
cycle for the steam engine may therefore be applied to
turbine calculations.
Figure 63 shows a portion of the ordinary temperature-
entropy diagram, with lines of constant pressure, constant
quality, and constant superheat. Three typical adiabatic
expansion paths are shown between the pressures of 80 and
30 lb.; Cidi starting and ending with wet steam, c-id-i starting
with superheated and ending with wet steam, and c^ds
starting and ending with superheated steam. The quality
of the steam at the end of the expansion (using the term
''quality" in a general sense as meaning either percentage
of dry steam in a wet mixture, or degrees of superheat when
above the temperature of saturation) is indicated by the
location of the points di, di, and d^ relative to the constant x
Unes or to the constant superheat lines of the chart.
The temperature-entropy chart, while inconvenient for
the designer and calculator largely because the value of
qa is represented by the difference of two areas, is very
useful for illustration of fundamental principles, and the
student should think as easily and naturally in terms of
this diagram as of any other. At least two other forms of
heat charts are preferable for calculation, however.
1 Figuratively; actually the work performed by each portion of expanding
steam is done on the steam immediately in front of it.
52 STEAM TURBINES
^f^
On the heat-content-entropy or MolHer chart, shown in

part m Fig. 64, the same three paths of adiabatic expansion
are shown, Cidi, 02^2, and 03^3. The available heats in the
three cases may
be obtained by taking the differences of
the ordinates of the points c and d on the vertical scale, or
by measuring the length of the line cd with dividers or a
strip of paper, and applying it directly to the scale.
Mollier charts are included with most texts on Ther-
modynamics and with all steam tables. Values vary
somewhat with different authorities, however, and when the
investigation now (1927) being conducted by the Steam
Tables Committee of the American Society of Mechanical
Engineers is concluded the result will be a table on which
full reliance may be placed. Until then, the chart pre-
pared by Dr. Keenan,^ based on a portion of the above-
mentioned investigation may be considered to represent the
best available data on the properties of steam. Reprints of
this paper, with the chart and table, are available for pur-
chase from the American Society of Mechanical Engineers.
The Mollier chart has the disadvantage of not giving
volume curves, the quality or superheat being read from
the chart, and reference made to a separate chart or table
of saturated and superheated volumes, with a multiplica-
tion operation in the case of wet steam.
The heat-content-volume chart prepared by Prof. F. 0.
Ellenwood- under the title "Steam Charts" is based on the
data contained in Marks' and Davis' Steam Tables, is
plotted in sections, and is bound in book form. All desired
data may be read directly from the charts without further
calculations, and the sectional form allows the variation in
scale from one portion of the chart to another that is
requisite for consistent accuracy. It is probably the most
convenient form of steam data for turbine calculations,
1 Keenan, J. H., "Progress Report on the Development of Steam Charts
and Tables, from Harvard Throttling Experiments," Mech. Eng., February,

1926.
2 Ellenwood, F. O., "Steam Charts and Special Tables for Turbine
Calculations," John Wiley & Sons, New York, 1917.
54 STEAM TURBINES
though it is not so suitable for representing the results of

such calculations as the Mollier chart which is generally
used for the purpose. In Fig. 65 is shown a portion of the
Ellenwood chart, with the same three expansions from 80
to 30 lb. as in Figs. 63 and 64.
9. Use of Steam Tables and Formulae.

If charts are unavailable, numerical steam tables may be
used, though the labor of calculation is increased. Special
pressure-entropy tables are very convenient in this respect,
though interpolation both directions may be necessary
in
unless the tables are extensive. ''Steam and Entropy
Tables," by Prof. C. H. Peabody,^ contains a very satis-
factory table of this sort, though the steam data used is
not the most recent. In ''Properties of Steam and
Ammonia," by Prof. G. A. Goodenough,- there is given a
small pressure-entropy table.
Finally, formulae derived from a study of the work of
various experimenters may
be used directly, insuring in
the results at least as high a degree of accuracy as our pres-
ent knowledge of the properties of steam allows. The
is great, however, and the degree of
labor of calculation
accuracy for engineering purposes unwarranted in view of
the present uncertainty as to the exact value of the various
coefficients which are used in the apphcations to turbine
calculation. If small drops in pressure are studied, how-
ever, use of these formulaemay be justified. The mathe-
matical expressions of the results being derived by the
A. S. M. E. investigation already mentioned are examples
of such formulae. "Properties of Steam and Ther-
modynamic Theory of Turbines," by H. L. Callendar,'^
is an elaborate investigation into the properties of steam
1 Peabody, C. H., "Steam and Entropy Tables," John Wiley & Sons,
New York City, 1912.
2 Goodenough, G. A., "Properties of Steam and Ammonia," John Wiley
& Sons, New York City, 1920.
3 Callendar, H. L., "Properties of Steam and Thermodynamic Theory
of Turbines." Longmans, Green & Co., New York, 1920.
and the application of the derived formulae to the design of

steam turbines. ''Steam Turbines," by W. J. Goudie,' con-
tains, in Chap. VI, a digest of the more useful of the Callendar
equations, with examples of their application, and in Chap.
VII, directions for the construction of special charts
intended to save labor in calculation.
Warning must here be given against taking tabular or
chart values from more than one authority and using them
in the same calculation. The values given by any one
authority are generally fairly consistent among themselves
over moderate ranges, even though the absolute values
may be in error, and trials will show that the final results of
successive calculations each consistently made with the
data of a single authority, will differ little among them-
selves. This will not be true, however, where high pressures
and superheats are involved, and in such cases the best
available data should be used.
10. Velocity of Flow and Quantity Discharged.

Returning to Eq. (4) (p. 50) and solving for V2,
= 223.8V/?7^/i7+~(77"-r223:8p. (5)
If the velocity of approach to the nozzle may be taken as

zero, as is ordinarily the case, Fi can be dropped, giving
F2 = 223.8V/11 - /i2. (5a)
Also, for continuity of flow, the following must hold at all

cross-sections,
Wv = AF,
whence the area A at any cross-section necessary to allow a
flow of W lb. per second of steam whose volume per pound is
V, is
A = -y. (6)
^ GouDiE, W. J., "Steam Turbines," Longmans Green & Co., New York,
1922.
.
56 STEAM TURBINES
Note that v is not the specific volume of dry saturated steam,

but of the steam as it exists at the particular nozzle section
under investigation, and it must be determined for each
case, from a chart or by calculation (see any text on
thermodynamics)
Applying Eq. (6) to the case of a constant flow of W
lb. per second through a nozzle, it is evident that the value
of A will depend on the ratio v -^ V. As the steam expands

through the nozzle toward a region of lower pressure, both
volume and velocity increase. The two rates of increase
will not necessarily be equal, and neither of them will be
constant, so that further investigation is necessary to deter-
mine the variation in A along the length of the nozzle.
11. Calculations for Nozzle with Ideal Frictionless Flow.
As an illustration, start with 1 lb. of steam at a pressure

of 100 lb. persquare inch absolute and a superheat of 120°
expanding adiabatically through a nozzle to a lower pressure
po = 90 lb. From': Ellenwood's Charts, Plate 2A, at
the intersection of the 100-lb. line with that of 120°
superheat, the value hi = 1,250 B.t.u. will be found, the
volume being 5.28 cu. ft. and the entropy 1.676.
Following this (interpolated) entropy line to the 90-lb.
pressure line, gives ho = 1,240 B.t.u., and volume = 5.72
cu. ft. The available heat ga = 1,250 - 1,240 = 10 B.t.u.:
the velocity V2 = per second (Eq. (5a) or Table V,
707 ft.
Ellenwood). Substituting this velocity in Eq. (6) gives A

1 V 72 '1
= ;: = 0.00809 sq. ft. = 1.164 sq. in. This is
for a flow of 1 lb. per second. For a different rate of flow

the area must be multiplied by the proper value of W.
the calculation be repeated, using a drop from 100 to
If
80 lb. it will be found that the area A = 0.881 sq. in. In
Table I are given the results of similar calculations made
with successively lower values of po- The column marked
''Diameter " gives the corresponding diameter of a circular
cross-section.
—
In Fig. 66 the values of /i2, volume, velocity, and area

are plotted on a base line along which the corresponding
values of p2 are laid out to scale: the total length of this
base line may be taken as the length of the nozzle. Using
the values of the diameter from Table I, the longitudinal
section of the corresponding nozzle may be drawn as shown
in Fig. 66a. A nozzle made with this profile would expand
1 lb. of steam per second from 100 to 2 lb., the pressure
drop along the nozzle axis being uniform.
Table I. Nozzle Calculations
Pi = 100; superheat = 120°; W = 1; <p = 1.6765

58 STEAM TURBINES
100 90 80 10 60 50 40 30 20 10 2 ^50 1200 1!S0 1100 1050 1000

Pressure p Heat Content hj
y
^ Fig. 67
F1966
V
/^
Fig. 66 a
Fig.67oi
ll
along the nozzle axis, and the diagram curves and nozzle
profiles would vary with each.
In Fig. 68 are shown the curves resulting if the nozzle
section. Fig. 68a, is made form to that in com-
similar in
mon use, having an entrance opening rounded with a circu-
lar arc leading to a flaringportion with a straight taper.
This form of nozzle is much used because it is easy to make,
and is probably at least as efficient as any other. It will
be noted that no function varies uniformly along the axis,
though the volume approaches uniform increase.
Minimum Section or Throat.
12.
However the data of Table I may be plotted, it will be

evident on inspection of any of the diagrams that as the
steam begins its flow along the nozzle length, its velocity
increases at a much more rapid rate than its volume.
Towards the outlet, however, the condition is reversed,
the volume increasing rapidly and the velocity slowly.
Applying these observations to Eq. (6), it is evident that
in the beginning, A
must decrease, while toward the end
it will increase. Somewhere between, there will be a mini-
mum area, v and V there having equal rates of increase.
The table indicates that this minimum is between 55 and
60 lb. That is, there is a converging entrance up to the
point where the pressure has fallen to this value, and
then a diverging portion for the rest of the length. The
minimum section is termed the throat.
13. Variation in Nozzle Area and Form, Constant Flow, Vari-

able P2.
The complete nozzle sections shown in Figs. 66a, 67a,

and 68a are designed to discharge 1 lb. of steam per second,
the initial pressure being 100 lb. and the pressure into which
the nozzle is discharging, 2 lb. If the discharge pressure
were raised to 5 lb., the nozzle should obviously bs cut off
at the point where the ordinate to the pressure curve is 5 lb.
as the expansion to 5 lb. is complete at that point. If p2
is raised to higher values, the nozzle will require correspond-

60 STEAM TURBINES
ing shortening; this is equivalent to saying that with

smaller drops of pressure, expansion is necessary.
less
When p2 = 58 lb.,expanding portion will have
all of the
been cut away, leaving only the convergent entrance, and if
P2 is raised still higher, the throat itself will be cut off.
Figures 66a, 67a, and 68a are thus each really composite
drawings of a series of nozzles, any one of which will dis-
charge 1 lb. of steam per second at 100-lb. pressure. Each,
however, is to be used with the back pressure appropriate
to the length of the particular nozzle chosen. The effect of
using a nozzle longer or shorter than appropriate for a given
po, termed " overexpansion " or ^'underexpansion," will be
discussed later in Sec. 25.

It is important to observe that as long as p2 is below 58
lb., the minimum diameter of the nozzle is constant what-
ever the actual value of p2, being that of the throat. It
also follows under these conditions that variations in p2
do not affect the rate of discharge. If p2 is increased above
58 lb., however, the minimum diameter increases above this
throat diameter. This means that with small differences
between pressures pi and p2, the nozzle is convergent.
14. Variation in Flow, Constant Area, and Variable po. Criti-

cal Pressure.
Viewing the subject from another angle, assume a divided

chamber, with a small opening in the central partition.
Let there be a supply of steam to chamber A adequate to
maintain a constant pressure of 100 lb., and assume that the
outlet C from B, connected to the atmosphere or to a con-
denser, can be manipulated so as to maintain any desired pres-
sure po. With C closed, there will be no flow, and p2 = Pi.
If C be partially opened, p2 will drop somewhat, and a flow of
W lb. of steam per second will be established. If this be
repeated, with decreasing values of p2, data will be obtained

from which a curve may be plotted, as in Fig. 70, showing
the values of W for different values of p2. The abscissae,
however, are values of P2/P1 = r instead of p2 directly.
FLOW OF STEAM IN NOZZLES
evident from Fig. 70 that as p2 is lowered, the flow

It is
of steam at first increases very rapidly, then more slowly,
and finally, when p2 is approximately 58 lb., a maximum
flow is reached, which is not
exceeded no matter how low the
final pressure is made. If the
pressure pi in A is 200 lb., the flow

at all values of p2 will be greater
than in the previous case, but a
diagram plotted from tests made
with this initial pressure would
show a curve similar to Fig. 70,
and the discharge would reach its
maximum value at about 116 lb., or
Fig. 69.
0.58 X pi. Similar results will be
found to hold with tests at other pressures, the flow always
reaching its maximum when P2/P1 = 0.58.
i.oo
^. 0.75
5 0.50
0.75
: :
62 STEAM TURBINES
accuracy, yet certain interesting facts appear and valuable

inferences may be drawn.
The equation representing the adiabatic expansion of a gas
is given by the following, k being the ratio of the specific heats
pv" = C, or pivl = povl. (7)
If this expansion be assumed to occur during flow through

a nozzle, the present problem resolves itself into finding the
value of p2, or, more generally, the value of the ratio r =
P'llpx for which the discharge is a maximum in the W
equation W= A
F2
—
The value of the final volume V2 is
V2
.
determined from Eq. (7) as
'=•.©'='# (8)
To find V2, rewrite Eq. (2) in terms of the properties of

a perfect gas
il + PlVl + 2g = ^"2
+ P2?^2 + 2fl'
^^^
where ii and 12 are the initial and final internal energies.

Dropping V\/2g for reasons already stated, and rearranging:
ll - l2 + PlVi - P2V2,
2g
PlVi P2V2
PlVi - P2V2,
K-l K-l
K
= ;^^r^(Piî - P2V2). (10)
/p-Y^
Substituting for P2V2 its value PivA—j from " the equa-
tion piVl = p2v\ and replacing p2/p\ by r will give

Vi
- Pfp)^-^\
K \ .
-^PxV^ «
2^^ J
-yPi^{ (r) (11)
Substituting the value of V2 from Eq, (8) and of Fo from Eq.

(11) in the equation W = AV2/V2 gives
"^ - ^^'^~il;(:' -^''} (12)

:
whence
W = A X Constant X \n - 7-'^. ^^^^
To find the value of the variable r for which W is a maxi-

mum, differentiate the quantity under the radical with
respect to r and equate to zero. Solving for r,
U= (^}^. (14)
The value of k is theoretically the same and = 1.41 for all

diatomic gases, but differs slightly for various actual gases
the pressure ratio Vq for maximum discharge could be found
for any particular gas by substituting the value of k for
that gas in Eq. (14). If, for example, the value k = 1.41
for air be inserted, then
ro = 0.527, for maximum flow of air. (15)

attempted to apply the equation to steam there
If it is
is some uncertainty as to the proper value of k to be used.
The equation
pv'-'^' = C (16)
due to Zeuner, has long been held to represent closely the
adiabatic expansion of steam initially dry and saturated,
apparently justifying the generally used value of 1.135 for n.
If this is substituted in Eq. (14), we have
To = 0.577, for maximum flow for dry saturated steam (16a)

For superheated steam, which, from molecular considera-
tions should have a value of k = 1.28, experience points to
pv'-' = C (17)
which, through Eq. (14), results in
To = 0.5457 for maximum flow of superheated steam (17a)
Comparing 16 and apparently a sudden change
17, there is
in the index when superheated steam expands

into the
wet region, which appears worth investigation. Various
considerations make it appear probable that steam thus
expanding into the 'Svet" region does not instantly partly
condense but remains for a brief time in an unstable
state termed "supersaturated," the index for this expansion
64 STEAM TURBINES
being the same as for superheated steam, namely, 1.3.
This whole subject will be considered later.
16. Critical Velocity for Flow of Perfect Gas.
Substituting the value of ro from Eq. (14) for r in Eq.

(11), the energy at the throat is
VI _ K
PlVi.
2g K - 1k + 1
Whence,
Substituting k = 1.41 for air, the throat velocity is
Vo = n.Wv^x (19)
Pi here being in pounds per square inch and Vi in cubic

feet per pound.
Another expression for the throat velocity may be
obtained by substituting for the initial conditions pi and Vi
in Eq. (18), the throat conditions, as follows: since povl =
V\^\ = Poôf
/poV^
— =
1
^-
I
Po?ô(ô)
v'"^'
-^ = yoô(
/« + 1)\
),
^
the last expression being obtained from Eq. (14). Sub-
stituting in Eq. (18),
Fo = VgKpoVo, (20)
which the same as the expression for the velocity of sound
is
in a perfect gas at the pressure and density existing at the

throat; Vo is therefore often called the acoustic velocity.
This is really the velocity with which a disturbance of
any kind, such as a change of pressure, will be propagated
through the gas. Taken in connection with Sec. 14, this
forms the basis of an ingenious though somewhat fanciful
explanation of the phenomenon of critical velocity, attrib-
uted to Prof. Osborne Reynolds. In Fig, 69, each time p2
is reduced, there is, in effect, a disturbance produced,
traveling with the velocity of sound in all directions:
in traveling upstream through the orifice, it affects the
F
conditions of the portion of the gas inA and more flow

results.When, however, the velocity through the orifice
has become equal to the velocity with which the ''news"
of the reduction of pressure travels, the gas in A is not
further affected by reduction of pressure in B, and the
velocity through the S7nallest section has reached its maximum,
or critical, velocity.
The significance of Eq. (12) in connection with the
critical pressure is illustrated in Fig. 71. With the initial
conditions pi = 100 lb. abs. and ti = 65°F., the values of

TT^ are calculated for values of r varying from unity to zero
1.25
Pr 100 lbs. a
%- 65°
6 1-00
.75
.50
^ .15
66 STEAM TURBINES
pi, po remains at the critical value, and it is the difference

between pi and po that determines the flow. The discharge
into a perfect vacuum cannot be less than the maximum,
and the portion W of the curve is to be discarded, T^^„
being used instead.
Velocities calculated by Eq. (11) may be similarly plotted
as ordinates on the base P2/P1, as in Fig. 72. The portion
V represents velocities for values of p^ between pi and
0.527 pi; the horizontal branch Vo shows the constant
velocity through the throat for all values of p2 less than po,
and the branch V2 shows the higher velocities that may be
zsoo
2000
ISOO
^1000
500
lb. pressure and 300° superheat. It is evident that varia-

tions of the critical velocity are rather limited.
68 STEAM TURBINES
but Eq. (23) for dry saturated steam

fairly reliable, is only
approximate; the method to be shown in Sec. 19 is
preferable.
19. Design of Steam Nozzles.

Returning to the steam nozzle calculations begun in
Sec. 11 and interrupted to establish the theoretical formulae
of flow for perfect gases for the sake of comparison, two
cases arise in steam nozzle design: A, small drops of pres-
sure resulting in convergent nozzles, and B, large drops of
pressure to exit pressures below the critical pressure, result-
ing in convergent-divergent nozzles.
A. Convergent Nozzle.
Suppose it is desired to calculate the dimensions of a
nozzle to expand steam from 130 lb. abs. pressure and 60°
superheat, to 105 lb. abs. The nozzle is to discharge 1.6
lb. of steam per second, and a loss of 1 per cent of the
available energy due to friction may be assumed. Calcu-

late the exit area and diameter.
Po = 130 X 0.577 = 75 lb., which is < 105 lb., therefore
the nozzle has no throat and is convergent.
From the Ellenwood charts,
Pi = 130; 60° superheat; h^ = 1,225.2; <^ = 1.622.

p, = 105; <p = 1.622; ho = 1,207.5 ; v. = 4.49.
Heat drop = Qa = 17.7 B.t.u.
V2 = 223.8V0.99 X "1777 = 937 ft. per second.

144 X 1.6 X 4.49 ,^^
A2 =
.
= 1.105 ,
sq. m.
^^
If round in section, ^2 = Vl.105 ^ 0.7854 = 1.186 in.
diameter.
If square in section, ^2 = \/l-105 = 1.051 in. square.
B. Convergent-divergent Nozzle.
Suppose that, in the preceding problem, the steam is to
be expanded to 16 lb. abs. pressure instead of 105 lb.,
that there is 1 per cent of energy loss up to the throat, and

10 per cent total loss. Find throat and exit dimensions.
Pi = 130; 60° superheat; hi - 1,225.2; <p = 1.622.

Po = 75; <p = 1.622; ho = 1,178.4; vo = 5.79.
Heat drop to throat = go = 46.8 B.t.u.

Vo = 223.8\/0.99 X 46.8 = 1,522 ft. per second.
144 X 1.6 X
> 5.79 ^-^
Aq =
.
r"roo ^ ^0.876 ^,
sq. m. throat area.
. ,
1,522
= 1.956 in. diameter (round).
= 0.936 in. square.
From the charts,
pi = 130; 60° superheat; hi = 1,225.2; ^ =1.622.

P2 = 16; <p = 1.622; h. = ^,065^; v^ = 22.55.
Heat drop = qa = 159.6 B.t.u.
V2 = 223.8x70.90 X 159.6 = 2,680 ft. per second.
144 X 1.6 X 22.55 . ^ô
A =
.
2 (T^^î^
"" 1.938 sq. m. exit area.
= 1.571 in. diameter (round).
The exit diameter is often rectangular when the throat

is square. If the height of the exit section be taken at 1.05
in., then the breadth = 1.938 -^ 1.05 = 1.845 in. Nozzles
of thisform are illustrated in Figs. 124 and 131.
In the above calculations the slight increase in volume
of the flowing steam due to heat arising from the friction
against the walls has been neglected : it will be considered in
Sec. 23.
20. Nozzle Friction.
Considering first the convergent portion of the nozzle,

the effects of reheating of the steam by friction, up to the
throat, are (a) to increase the entropy of the steam, (6) to
diminish the actual net heat drop for any given pressure
limits, (c) to diminish the velocity of flow at the throat,
(d) to increase the volume of the steam, and (e) to increase

the throat area necessary to discharge a given quantity of
70 STEAM TURBINES
steam, or conversely, to decrease the discharge for a given

throat area.
For the well-formed round-section nozzles used in exper-
iments, these effects are all so slight that they may be neg-
lected. For actual turbine nozzles which are usually square
or rectangular in section for the sake of increased efficiency
of action on the blading, and which are usually curved,
and longer than necessary from expansion requirements
alone, the losses up to the throat are not negligible.
Friction beyond the throat will not appreciably affect
the discharge quantity, but has the same effects otherwise
as mentioned in connection with throat friction.
Frictional losses are less with (a) large nozzles, in which
the cross-sectional area divided by the perimeter (hydraulic
mean radius) is larger, (6) round cross-section as compared
with elliptical, (c) moderately
square, or rectangular section,
which allow a shorter nozzle for a
large angles of divergence,
given ratio of expansion, and (d) superheated steam as
compared with dry or wet steam.
21. Nozzle Efficiency.
The efficiency of a nozzle is always less than unity because
of the effect of friction, which is always to destroy some of

the kinetic energy, returning its equivalent in heat energy
to the flowing stream at a lower temperature than that
originally by it. This irreversible operation
possessed
increases the entropy and a portion of the original energy
is rendered unavailable. This is termed reheat, is measured
in B.t.u., and is represented by Qr.
Nozzle efficiency e„ is the ratio of the actual kinetic energy
of the steam at exit, to the theoretical kinetic energy.
If V'2 is the actual velocity at exit and V2 is the theoretical
velocity, then Fo = knV2 where kn is called the velocity
coefficient.
Also,
AVJ
2(7
_^
'
AVI
2g m '
«
whence,
"2
Vt= vu
and
F; = F2\/e„ = 22S.SVeJh, - h^) = 223.8Venqa'- (24)
where
hi — h-2 = g„ = heat energy available for producing velocity.
The expression e„(/ii — ^^2) = ^nqa, which is the heat actually
used in producing velocity, will be called d = net heat drop.
Also, since
Available energy = net heat drop + reheat,
or
Qa = ei -\- Qr = enQa + Qr,
then
Qr = qa — enQa = Q a{l " e„). (25)
If the initial velocity Vi is appreciable,
energy theoretically available = hi — h-z + ( o9o"o ) >
and
equivalent net heat drop = €„ qa + ( ooTs) ' ^^^^
This assumes that the same efficiency applies to both por-

tions of energy. Also,
reheat = q,. = (I — e„) qa + (27)

V223.8/
This reheat must be added to ho, the heat that would remain
after true adiabatic expansion, to find the actual heat
content of the steam at exit = h'o,
ho = actual heat content at exit = Ao + qr. (28)

From the heat content at exit and the exit pressure, the
volume ^2 and entropy at exit may be found from charts
or tables, thus completely determining the steam condition
at exit.
22. Nozzle Experiments.
Accurate knowledge of all the phenomena accompanying

the flow of steam through nozzles and passages is essential
:
72 STEAM TURBINES
to good turbine design and much experimental work has

been carried on by scientists in many countries. The data
most desired have been:
A. Discharge in pounds per second per square inch
nozzle area (leading to values of the discharge coefficient c)
for various initialand final pressures and for various sizes,

forms, materials, and finishes of nozzles. The usual method
is to condense and weigh the issuing steam. Values of c
obtained with a straight or moderately curved round-
may run from 0.97
section nozzle with well-rounded inlet
to 1.00.
B. Velocity coefficient kn = ratio of actual discharge
velocity to theoretical discharge velocity, the square of
which is usually assumed to be the nozzle efficiency €„.
If the velocity not uniform over the cross-section of the

is
jet, the efficiency thus calculated will be slightly low,

since XmV" < Fat, 2m.
To find the velocity coefficient and efficiency, the follow-

ing methods have been used
a. Measurement of weight discharged and estimate of
density at point of exit, allowing calculation of velocity.
This method is inaccurate because experiments have
definitely shown that high coefficient of discharge is not by
any means necessarily associated with high efficiency.
The temperature and quality at exit are also uncertain.
b. Direct measurement of velocity of exit by some form
of impact tube, such as the Pitot tube.^ This is a very

accurate method for exploring the whole cross-section of
the discharge of a nozzle with an impact tube, using air,
however, instead of steam. Accurate knowledge of the
distribution of velocities over the section is obtained, so
that the assumption in B above need not be made. The
effect on the nozzle flow of the ''preliminary conditions,"
meaning the treatment the air stream has received before
1 An apparatus vising this method is described in an article entitled "An
Experimental Investigation of Nozzle Efficiency," by H. Loring Wirt,

published in the Trans. Amer. Soc. Mech. Eng., 1924, p. 981.
.
it enters the nozzle inlet, is especially well studied by this

method (Fig. 73 illustrates the principle)
c. Provision of a movable or flexible support for the nozzle
and the chamber from which it discharges, so that the

reaction of the jet leaving the nozzle may be measured,^ as
indicated in Fig. 74; from this and the weight discharged,
the velocity can be calculated
by the equation V = Fg/W.
d.jMeasurement of the force
exerted on a flat plate when the
jet is allowed to strike normally
at its center, the velocity being
calculated by the same formula
as above, V = Fg/W.
The most extensive, com-
prehensive, and accurate series
of experiments so far made
by
public are those carried out
method (d) by
Steam
Nozzles Research Committee
the
\\\\\U
~Pre/im/nary guides
of the Institution of Mechan-
ical Engineers (British).- Fig-
ure 75 shows the form of appa-
ratus used. The data obtained
I Air supply
and conclusions drawn are

Fig. "3. — Nozzle-testing
machine.
Impact-tube method.
too extensivebe even to
abstracted here, but the reading of the reports and
discussions will amply repay the interested student.
Unfortunately, the pressures were limited to the range
between pi = 40 lb. gage, and atmospheric pressure.
Convergent or parallel nozzles made from reaction form
blading, as well as impulse nozzles, with both thick and thin
nozzle partitions of uniform thickness were tested. Ques-
1 Warren, G. B., and J. H. Keenan, "An Experimental Investigation of
Nozzle Efficiency," describes a recently devised apparatus working on this

principle, Trans. Amer. Soc. Mech. Eng., p. 33, 1926.
^ Reports of the Steam Nozzles Research Committee, Proc. Lnst. Mech.
Eng. (British), pp. 1 and 113, 1923; pp. 445 and 715, 1924; p. 547, 1925.
:
74 STEAM TURBINES
tions of supersaturation (see Sec. 29) were avoided by
working entirely in the superheated region.
The committee expressed most of its results in the form
of curves showing the variation in velocity coefficients with
different values of the theoretical exit velocity. While
the form and position of the curves were affected by numer-
r-i%" Overall
Fig. 74. — Nozzle-testing machine. Reaction method.
ous conditions, some still obscure, the following conclusions

appear to have been well established by the work of the
committee
(a) Between theoretical exit velocities 700 to 900 ft. and
1,700 to 1,800 ft. the coefficient was substantially constant

varying from 0.90 for an evidently poor form of nozzle,
to 0.96 for a very good nozzle, with most of the nozzles
having values ranging from 0.93 to 0.95 (corresponding to
efficiencies of 0.865 to 0.901). The coefficient for an
"ideal" straight nozzle of round section for the same range

was about 0.98.
(b) At lower velocities, down to the lowest obtained,
about 400 ft. per second, there was a notable increase of

from 3 to 5 per cent in the coefficient.
(c) At higher velocities, up to about 2,100 ft. per second,
the highest obtained, the

coefficients rose slightly
with some nozzles and fell
with others.
The solid portion of the
curve in Fig. 76 represents
these results. The dotted
portion to the left is an
extension proposed by H.
M. Martin in line with
known facts concerning
slow (viscous) flow, high-
velocity (turbulent) flow,
and moderate-v e 1 o c i t y
(mixed) flow. If this is
true, then very low veloc-
ities (less than in turbine
practice) may be accom-

panied by low-velocity
coefficients.
Warren and Keenan,
however, in their article -Nozzle-testing machine,
pulse-plate method.
do not agree to all these
conclusions. Results of their tests on a straight, round,
converging nozzle and an impulse nozzle are shown in
Figs. 77 and 78, respectively. They agree that the
velocity coefficient shows evidence of rising slightly beyond
the velocity of sound, but then falls off rapidly with
further increase, up to 2,500 ft. per second and beyond.
They find no reliable evidence, however, of the rise at low
velocities, and their values for both velocity coefficients
76 STEAM TURBINES
and discharge coefficients are notably higher than the

British results.^
1.00
.98
:.96
.94
.9? t Ve/oc/fy^ coeff.
.90 14-
Fig. 76.- — Summary for impulse turbines. {Steam Nozzles Committee.)
Fig. 77. — Converging nozzle, 1.6 in. diameter. (Warren and Kcenan.)
1.0
C.98
•".96
Ve/oc/fy^ coeff
.94
500 1000 1500 2000 ?500 5000
Theoretical Velocity Ft. per Sec.
Fig. 78. — Impulse turbine nozzle. {Warren and Keenan.)
Passing to the methods of estimating nozzle losses

given by other authorities, Callendar gives the follow-
1 Those who wish to pursue this subject further should also read Chap. II
of "Steam and Gas Turbines," by A. Stodola, trans, by L. C. Loewenstein,
McGraw-Hill Book Company, Inc., 1927.
:
ing as being fairly representative of older experimental

results
Or = reheat in B.t.u. to throat section = , (/ii — /12),

Pido
(29)
where do is the throat diameter. For 100 lb. pressure and

1 in. throat, this would amount to 1 per cent loss; for ji in.
diameter, 4 per cent loss.
Theoretical Velocit^/=V2
1000 2000 3000 4000

1.0
^
90
80
.70
78 STEAM TURBINES
value, or, in other words, for converging nozzles. It is

probably not reliable for heat drops of less than 100 B.t.u.
The curve A (Fig. 79) shows average velocity coeffi-
cients for various heat drops according to one American
manufacturer, applying to square or rectangular nozzles
of the form used in impulse turbines and of such an area
at exit that expansion is just complete. The coefficient
falls off continuously with increase of adiabatic heat drop
in this diagram, w^hich fact is not in accordance with the

results obtained by the Steam Nozzle Committee.
Curve B shows the same data, plotted with values of
efficiencies as ordinates.
If the nozzle exit is too small or too large, giving under- or
overexpansion, there is a loss of efficiency which will be
discussed in Sec. 25, and the coefficients from Fig. 79, or
other data, must be multiplied by the factors there given,
in order to get the true velocity of discharge from such a
modified nozzle.
In reaction turbines, the steam undergoes expansion in
both moving and fixed blading, both of which, therefore,
act as nozzles, and are subject to the above discussed losses.
These will be discussed in Sees. 51 and 52.
23. Effect of Reheat on Nozzle Design.

In Eq. (28) it was shown that h'o = h-i + g,. To illustrate
the effect of this additional heat in slightly increasing the
volume at exit and hence the necessary nozzle area, the calcu-
lation of part B, Sec. 19 will be continued. From Eq. (25),
q^ = g„(i _ e„) = 159.6(1 - 0.90) = 15.96 or 16 B.t.u.

h2 = 1,065.6 -h 16 = 1,081.6 B.t.u.
From the chart, at this heat content and with po = 16 lb.,
v, = 22.60 cu. ft. instead of 22.55 cu. ft.
^^^ ^^'^
The corrected area = A', = ^^'1^ = 1-943 sq.
in. area instead of 1.938 sq. in.
Since, in practice, nozzles are nearly aways made shorter

and, hence, smaller in exit area than theory would indicate
desirable, there is little error in neglecting this increase in

volume due to friction, unless the nozzle efficiency is very
low indeed.
24. Condition Curve.
The increase in entropy of the steam expanding through

a nozzle may be represented on any of the steam charts.
In Figs. 63, 64, and 65, the point marked di indicates the
state of this discharge steam, and the curve C2 — d'i repre-
sents what is called the condition curve of the expansion,
or sometimes simply the expansion curve.
The exact form of this curve for a single nozzle is of no
particular consequence, but in the case of a turbine involving
a sequence of similar expansions extending over a much
greater pressure range, the curve form is important and
will be considered later (curve c-zd'ê-i, Fig. 64).
25. Over- and Underexpansion.

In the actual operation of turbines, nozzles are often sup-
plied with steam at a different pressure from that for which
they were designed. If a nozzle in a turbine governed by
throttling is designed for correct expansion under full load,
then at part load pi and p2 will both be decreased, pi more
than P2, and the ratio r = p2 ^ pi will generally be more
than that for which the nozzle was designed: the nozzle is
then overexpanding. The same condition is brought about
by an increase in p2 the absolute discharge pressure (a
decrease in vacuum). Again, under certain conditions,
a nozzle may be supplied with steam at a higher pressure
than designed or p2 may be decreased, leading to a condition
of underexpanding.
Experiments have been conducted on nozzles in which
the pressures at various points were measured by gauges
connected to small holes drilled in the nozzle walls, or by
a single gage connected to a movable "searching tube"
passed along the axis of the nozzle, with a tiny hole in its
wall (Fig. 806). Various forms of nozzles have been
:
80 STEAM TURBINES
investigated,and while the results are not all in exact

agreement, the following facts are clearly indicated for a
convergent-divergent nozzle
Distance Along Noixle Axis

Fig. 80a.
Search inpf tube
Fig. 805.
When p2
A. is well above the critical pressure, the pres-
sure drops sharply at the throat and rises again beyond it
(Fig. 80a, curves A and B). Evidently, in addition to
expanding the steam, the contracted throat has a venturi-
tube effect, increasing the velocity and decreasing the
pressure, part of the latter being regained. There are often
zones of high pressure and low pressure alternating along
the flaring part, indicating an alternate compression and
expansion, gradually smoothing out as the nozzle exit is
reached (curve B).
2. When p2 is less than the critical pressure, but still
above the proper exit pressure for the particular nozzle

used, the expansion in the nozzle beyond the throat drops
the pressure to a point below po. It then rises and falls
more or less regularly as flow continues toward the exit
(curves C and D).
3. As p2 approaches the pressure which the nozzle is
for
designed, this wavelike compression and expansion gradu-
ally disappears, resulting in a smooth regular drop in pres-
sure throughout the length of the nozzle (curve F).
20
82 STEAM TURBINES
It is evident that there must be losses when either over- or

underexpansion occurs, and it is of importance to find how
hirge these losses are, both absolutely and relatively.
Figure 81 gives curves derived from data in a paper by
C. P. Steinmetz and said to represent these losses under
average practice conditions. The full lines represent the
loss of velocity in per cent, and the dotted curves the loss of
energy. It is evident that a deficiency in exit area (causing
underexpansion of, say, 5 per cent), causes much less energy
or velocity loss than the same per cent excess in exit area.
If, therefore, nozzle ratios of expansion must be varied
in turbine operation, the nozzles should be so proportioned

that considerable underexpansion takes place under the
greater loads, for the sake of avoiding the excessively
wasteful overexpansion at part loads.
26. Divergence of Nozzle Walls.
If is of the simple reamed type with straight

the nozzle
from throat to exit, the divergence should not
conical flare
be so great that the expanding jet tends to lose contact
^\ with the walls, as experi-
ment shows that this causes
considerable loss. Experi-
ment also seems to indicate,
however, that the angle of
divergence may be as great
Fi*^- ^^-
as 20 deg. without appreci-
able effect on nozzle efficiency, probably because the
shorter nozzle resulting from more rapid divergence
decreases friction on the surface. A jet so widely diverg-
ing cannot be efficiently directed into the blades, how-
ever, and a compromise is found in the common use of
angles varying from 6 to 12 deg. (total included angle).
An angle of 15 deg. is recommended by one manufacturer
for condensing turbines, and 10 deg. for non-condensing
turbines. Figure 82 shows a simple form of nozzle with a
straight flare as used by one company. Occasionally a
parallel extension is formed beyond the divergent part, as in

Fig. 83; while this directs the jets more accurately, it
increases the loss due to nozzle friction and is more expen-

sive to manufacture, so that its desirability
is a matter to
be settled by careful tests and commercial considerations.

Experience shows that the entrance to a nozzle should
be well rounded, so that there shall be no contraction
beyond the throat.
Turbine nozzles are always cut off on a slant as shown in
Figs. 82 and 83, in accordance with the angle a made by
the axis with the plane of blade rotation. The nozzle
Fig. 83.
length in Fig. 82 is measured as shown and it is apparent

that the steam leaving the nozzle on the side a is under-
expanded, and on side b is overexpanded.
27. Angle of Nozzle Position.

The angle a, made by the axis of the nozzle with the plane
of rotation of the blading, is determined mainly as a com-
promise between conflicting practical considerations.
Higher theoretical efficiency may be obtained by use of
small nozzle angles, as the blade action may then be
more efficient. This, however, will make the necessary
length of the nozzle so great that its frictional loss is greatly
increased, the total deviation of the steam in the blading is
increased, again causing frictional loss, and the diagonally
cut mouth of the nozzle is so wide that it fails to guide the
84 STEAM TURBINES
steam properly into the blading. If the angle a is made too

great, the blading loss becomes large, as will be shown in
Chap. III. The values used in practice range from 10 to
30 deg., 12 to 20 deg. being common.
28. Nozzle Forms.
The form of nozzle, illustrated in Figs. 82 and 83, is

used only in the case of single-stage impulse turbines or
in the high-pressure stage of some pressure-compound
impulse turbines. Instead of being round in section,
Fig. 85
nozzles may
be square or rectangular, this being partic-
ularly advantageous when a series of nozzles is arranged in
a row with the idea of delivering a solid belt of steam onto
the blading. The first high-pressure nozzles are also gen-
erally formed in a part called the nozzle block or nozzle
plate. Figure 84 shows a nozzle block carrying a series of
reamed nozzles and Fig. 85, a block with rectangular sec-
tion filed nozzles. In both of these steam is flowing toward
the observer and to the right. The increased nozzle loss in
Fig. 85, due to the greater perimeter of the square section of
each nozzle, is more than counterbalanced by the greater
efficiency of the jet action on the blading.
Working these square-section nozzles out of the solid,
even though they are roughly cast to shape, is exceedingly
expensive. With the exception of the first-stage nozzles

of some multiple-stage impulse turbines which receive
steam at high pressure and superheat and which are few
in number, the nozzles are usually formed by special parti-
tions or blades set (generally cast) into rings fixed in the
casing, receiving steam on one side and delivering it to the

blading on the other. Figure 86 shows such an arrange-
ment of orifice plates, and the dotted lines indicate how they
may be thickened along the middle in cases where it is
thought the blades need strengthening. This gives them a
Fig. 87.
form resembling stationary reaction blading. Figure 87

shows the plates set in a casting, a portion being broken
away to show their form better.
These discs in which orifice plates are set are termed dia-
phragms, and, taken in pairs, form the chambers in which
the various wheels rotate, each chamber having a different
86 STEAM TURBINES
pressure. These diaphragms are indicated by (18) and

(21) in Fig. 19.In the earlier stages the sections of nozzle
blading extend only partially around the circumference,
giving what is termed partial peripheral admission.
Toward the intermediate stages, the increasing volume of
the steam necessitates increasing the arc occupied by these
groups, until they eventually fill the whole circumference,
except for the necessary blank spaces on the horizontal divi-
sion Hne. Thereafter increased nozzle area must be obtained
mainly by increasing the height of the nozzles, often com-
bined with moderate progressive increase in the diameter
of the blade ring and, in the last few stages, by increase of
the nozzle exit angle (with plane of rotation). Figure 88
Fig. 88. — Impulse turbine diaphragms.

shows a diaphragms, illustrating the nozzle
series of these
increase inboth circumferential and radial directions.
In Fig. 90 is shown a half-section of a diaphragm, the
nozzle passage being shown at the top, and receiving flow
from the left. The turbine shaft revolves within the cir-
cular hole left at the center, and leakage of steam at this
point is stopped by what is termed labyrinth packing.
This is better shown in Fig. 89.
Diaphragms and wheels follow one another in close suc-
cession, and some attempt is nearly always made to con-
serve the exit velocity from a preceding wheel by forming
the nozzle entrance for the succeeding stage diaphragm in
such a way that it will receive this steam entering (with a
velocity which has previously been called Vi) and turn i
in the direction required for delivery into the moving
blading at the velocity V2. Diaphragm nozzles, formed by
partition plates, are therefore always curved and conse-
quently have greater length and somewhat more friction

than straight nozzles. Figure 86 illustrates this point.
The nozzle blocks of the first stage, which receive the steam
of highest temperature and pressure, are usually made of
phosphor bronze, nickel bronze being sometimes substituted
for pressuresand temperatures above 250 lb. and 600°.
Fig. 89. —Diaphragm and shaft packing. Fiu. 90.

(General Electric Company.)
In the intermediate and lower stages, the nozzle division

platesmay be of nickel steel pressed to form, or of lengths
cut from rolled bars of the proper section. The stationary
nozzle blades are, of course, not subject to centrifugal
stresses,hence freedom from corrosion and erosion, and low
cost of manufacture, are more important than strength.
The General Electric Company has developed a system
of separate nozzle rings or plates which may be threaded
88 STEAM TURBINES
into dovetailed grooves turned in the diaphragm face.

The rings may then be made of a different material from the
diaphragm body, and may be cast and machined inde-
pendently of the latter at a saving in cost (Fig. 90).
29. Supersaturation.
Equation (23) for the flow through a nozzle of stream

initially dry saturated is, with the addition of a coefficient
of discharge c,
W = 0.3044 Ac Vpi H- vi. (23)
In its derivation, Zeuner's equation pv^'^^'^ = C was used,

and it was assumed that the steam acted like a perfect gas.
Investigators early noticed that the results of actual
tests of nozzles gave results which would
on the discharge
make very close to unity in the above equation, and in
c
some well-authenticated tests, even greater. This cannot

be explained by the usual theory, since various conditions
always cause a reduction of the coefficient of discharge
below the ideal value. It is now known, however, from
experiments and research by H L. Callendar, C. T. R.
.
Wilson, and others, that dry-saturated steam expanding

suddenly as through a nozzle, instead of becoming slightly
wet immediately and progressively wetter as the expansion
proceeds, as was the case in the experiments from which the
exponent k = 1.135 was deduced, will not start to condense
until a certain lower pressure has been reached, and until
then remains in an unstable state, apparently vaporous,
termed supersaturated or undercooled. Apparently there
is necessity for nuclei or centers around which the mole-
cules can cluster during condensation, and form droplets.

The time element is also involved, and even if nuclei (such
as dust or particles of moisture) are present, the extremely
brief time for passage from the initial to the critical condi-
tions,being of the order of magnitude of one ten-thousandth
of a second, will prevent the establishment of complete
equilibrium.
Down to the pressure (different for each initial pressure)

which appears to Umit the amount of supersaturation,
called the Wilson limit, the adiabatically expanding steam
acts like a gas instead of condensing and the equation pv''
= C can therefore be used both experimental investigation
;
and theoretical analysis point to the value k = 1.3 as proper

for both superheated and supersaturated steam expanding
under orifice and nozzle conditions.
It may be said then that Eq. (22), namely,
W = 0.3155 Ac Vpi ^ vi,
with suitable values of discharge coefficient c, will ade-

quately represent the flow of superheated or saturated
steam when the pressure ratio r is less than Vo = 0.5457,
units being square inches as before. For pressure ratios
above the critical, Eq. (12) applies with the substitution of
K = 1.3 and the introduction of the actual value of r.
30. Supersaturation Limit.
Callendar states that, for steam initially saturated, ''the

limit of the supersaturated state is reached when the
pressure has fallen to about one-third of the initial value."
This Wilson limit is not yet well defined, but it always
appears to be well below the critical pressure and it is
therefore accepted by some authorities that the steam
remains supersaturated until after the throat section is
passed, and then regains equilibrium by condensing more
or less rapidly.^ A line roughly representing this limit
drawn on the Mollier chart would extend from the inter-
section of the 200-lb. line with the 93 per cent quality line,
to the intersection of the 0.5-lb. line with 96.5 per cent
quality. Down to this line, the approximately straight
constant pressure lines in the wet region will be replaced
by continuations of the curved constant pressure lines from
1Martin, H. M., "A New Theory of the Steam Turbine," Engineering,
London; a seriesof articles commencing July 5, 1918, advances the proposition
that equilibrium never complete, supersaturation continuing throughout
is
the turbine even into the exhaust. Study of this article should be deferred
by the student until after he is fairly familiar with the ordinary theories.
:
90 STEAM TURBINES
the superheated region. Constant quahty Hnes will be
replaced by lines parallel to the saturation line called lines of
constant undercooling, marked in degrees like lines of con-
stant superheat. As there is no change of state, the satura-
tion line will lose much of its significance, becoming merely
the dividing line between superheating and undercooling. ^
The following adiabatic relations will evidently apply

under conditions of supersaturation
1 GouDiE, W.
s-(;r-s =(;)'
J., "Steam Turbines," Longmans, Green & Co., p. 223, ed.
1922, the author discusses supersatviration and also gives a folded Mollier
chart constructed in accordance with the above scheme.
As an example, suppose that at a certain stage in a

turbine dry steam at 100 lb. expands through a nozzle
abs.
to 40 lb. abs., which is not far from the Wilson limit. Using
Ellenwood charts, hi = 1,186.5, î = 327.8, T, = 787.8, Vi =
4.43. Since undercooling takes place, the adiabatic equation
/»l\ ^ /100\0.769
applies, and V2 = v^ (—)'' = 4.43(^^1 = 8.96 cu. ft.
This than 9.89 cu. ft. which would have been the
is less
volume if thermal equilibrium has been maintained during

the expansion coupled with condensation of part of the
steam, giving x = 0.943. Also To = 787.8 (4%oo)^'' =
637.6°abs. = 177.6°F. If the steam had expanded in a
condition of equilibrium, the final temperature would have
been 267.3°F.; the undercooling is 267.3 - 177.6 = 89.7°F.
In Fig. 91 shown a portion of the Mollier chart illus-
is
trating the problem. The adiabatic starting at point a

intersects the 40-lb. supersaturated line at b. This being
assumed the supersaturation limit, re-establishment of
equilibrium takes place at constant heat content, resulting
in the condition shown by point c on the same horizontal
line as b but on the 40-lb. line for wet steam. The heat
drop is the vertical distance between the points a and c,
whereas, if equilibrium had been continuously maintained

during the expansion, it would have been the amount ae.
Thus, the effect of supersaturation in diminishing the net
heat drop is similar to friction.
The effects of taking account of supersaturation in calcu-
lations are, to increase the discharge for a given nozzle
about 5 per cent over what would be expected with the
usual method of calculation, and to increase the entropy and
volume after the Wilson limit is passed and equilibrium
established. In usual American practice no separate
account is usually taken of supersaturation, its existence
simply modifying slightly various coefficients used in design.
CHAPTER III
BLADING
The purpose of the blading in a turbine is to receive the

steam directed into it from the nozzles, and, by means of
the curved shape of the blade surfaces, to transform as
much as possible of the kinetic energy of the incoming steam
into useful work, evidenced by the force exerted on the
blades in the direction of motion. Ideally perfect blading
would offer no frictional resistance to the steam flow,
would allow streamline flow^ with no eddying, and would
absorb all of the kinetic energy from the steam, leaving
zero exit velocity. None of these conditions can be ful-
filled completely, and as a consequence the efficiency of the
blading is always less than unity.
31. Elementary Forms of Blading, Impulse.
Before developing the relations existing in the forms of

it is desirable to consider some of the
blades in actual use,
Fig. 95.
fundamental principles as applied to simplified forms. If

a smooth stream of water is deUvered from a nozzle directly
against a flat plate, as in Fig. 95, it will spread out in all
directions over the plate and will exert a steady force on
the plate, urging it in the direction of the jet, which force is
called an impulse. The water has its velocity V2 in the

92
BLADING 93
horizontal direction entirely destroyed; if W pounds of

water are discharged per second the force or impulse F will
be WV^/g. No work will be done, since the plate is
assumed to be at rest.
If, as in Fig. 96, the plate is moving toward the right (in
the direction of the motion of the jet) with a velocity Vh the

jet strikes the plate with the velocity V2 — Vh. This is
the velocity of the jet relative to the plate or, as it is usu-
ally termed, the relative velocity. The resulting impulse
(steady force) acting on the plate is then F =
W (Fo
— — Vh)
W
and the work done = — (F2 — Vb)Vh. Note that the work
is zero when F^ = 0, also when Vb = Fo, the latter being
true because the impulse is then zero. To find the value of
Vb (between and F2) at
which the work and effi-
ciency will both be maxima,
diff eren t i ate the expres-
sion for the work with
respect to Vb and equate ^^^ g^
to zero. Fo - 2Vb = 0.
Solving, Vh = V-2/2, or the velocity of the plate should be
half that of the jet. With this velocity, the amount of work
done IS — I F2 ^ )^ = -7 — which is evidently halt
the original kinetic energy of the jet, and the efficiency, the
maximum obtainable with this arrangement, is 50 per cent.
If, instead of a flat plate, the jet be directed on a cylin-
drical blade so shaped and placed as to turn the jet through
180 deg. as in Fig. 97, and if this blade be fixed in position,
the impulse will be twice as great as in Fig. 96. If the
blade is moving in the direction of the jet with a velocity

Vb, the velocity with which the jet enters the blade, or,
briefly, is Fo — Vh, and, neglecting
the relative velocity,
friction, be the relative velocity at exit. This
this will
will be directed oppositely from the original jet, however,
94 STEAM TURBINES
SO that the absolute velocity of the jet at exit will be Vb —

(F2 — Vb) = 2Vb — F2. The total change in velocity
will then be V2 - (2F5 - F2) = 2(72 - Vb), and the
impulse F = is
W2(^2
— — Vb). The work done per second
2W,.,, l^ 2WVbVJ^ Vb\ 2WVIp.. ,
where p = Vb/V2 = ratio of blade speed to steam jet

velocity. To find the conditions for maximum work,
differentiate wdth respect to p and equate to zero as before
2Fi 4Fip ^ ^
9 9
Solving, p = }'2> showing that in this case also the blade
speed should be half the jet velocity for maximum efficiency.
Substituting this value of p, the work done is found to
be WVl/2g, which is the total kinetic energy of the jeii, and
2V9
the absolute velocity of the jet leaving the blade is —^
— F2 = 0, whence the ideal efficiency of the blade is 100
per cent. In general, with the type of blade considered,
the efficiency is
2F1p,
(1 - P)
= Q = -
eii 4(p p2). (32)
Vy2g
32. Reaction.
Referring to Fig. 95, a force is evidently necessary to

give the acceleration to the water as it flows through the
nozzle. This force, as already stated, is WV2/9 and there
is a corresponding and equal force tending to move the
vessel from which the water is flowing, in the opposite direc-

tion, this force being termed reaction. Other examples
of reaction are (a) a rocket, in which a fluid is produced within
the moving part and expelled at high velocity, this velocity
generated within the moving part being the cause of the
reaction; (b) a whirling lawn sprinkler, where water is led
into the moving arm along the axis of rotation and there-
BLADING 95
fore exerts no turning effort at entrance, but only when

leaving at right angles to the arm; and (c) a jet entering a
cylindrical blade as described in connection with Fig. 97.
The last case merits analysis; the action may be divided
into two parts; while the being turned through 90
jet is
deg., the blade receives force toward the right due to impulse.
While the jet is being turned through the second 90 deg.,
we may say that the velocity Fo — Vb toward the left is
Fig. 98. Fig. 99.
being generated within the moving blade, and, the jet being
directed toward the left, there is a reaction exerted on the
blade, urging it toward the right. The action of the jet
on the cylindrical blade is therefore a combination of
impulse and reaction, though in steam turbines it is gen-
erally termed impulse, and the term reaction is reserved for
an action more nearly resembling that of the rocket. In
the case of the turbine, the generation of velocity is due to
the expansive action of the steam, which can leave a cham-
ber or passage at a greater velocity than it had at entrance,
and thus cause reaction on the chamber.
33. Graphical Representation of Impulse-blading Velocity

Relations.
In the case of the cylindrical blade, indicated above,

in Fig. 97, relations at entrance to the blade are shown in
:
96 STEAM TURBINES
Fig. 100a, and at exit in Fig. 1006. The following nomen-
clature will be used hereafter
V2r= relative velocity at entrance = V2 — Vb.

Vsr= relative velocity at exit = — (F2 — Vb).
F3 = absolute velocity at exit = Vb - (Vo - W). (33)
Note that V2r = F3,. numerically only under ideal condi-

tions, when friction is neglected.
There is no successful turbine
l/l
in which the blades move directly

away from the nozzle as in Fig.
97. Figure 101 shows the prin-
ciple of the usual arrangement,
and Fig. 102 one type of vector
diagram used for studying the
The blade velocity Vb is
action.
shown to scale by cd and the veloc-
ity of the jet leaving the nozzle by ad, the angle between
them often being 20 deg. The velocity of the stream rela-
tive to the blade is therefore ac, marked F2,- The angle of
the surface of the blade at the point where the stream
BLADING 97
velocity as at entrance. Let ce = F3,. represent this

relative exit velocity. While the stream is leaving the
blade at this velocity the blade itself has the absolute
velocity ef = The stream at exit, therefore, has the
Vb.
absolute velocity cf, marked F3.
A little reflection will show tht Fig. 100a is the limiting
case of the triangle acd of Fig. 102 as the nozzle angle is
decreased to zero, also Fig. 1006 is the limiting case for the
exit triangle cef of Fig. 102 as the blade-exit angle is
decreased. A good grasp of the relative and absolute
velocity relations of Figs. 101 and 102 is essential. It will
98 STEAM TURBINES
ular to the direction of Vb] V^j and F3/ are such compo-
nents V2 and F3, respectively, and are called axial
of
components, or velocities of flow, at entrance and exit, respec-
tively. The angle made by the entering jet with the line
of motion of the blade, often called the nozzle angle, is
marked a; that made by the relative velocity with the

same line is /?; the relative exit angle 7 and the absolute exit
angle 5 are as marked. In Fig. 104 is shown a similar
diagram, but with a much lafger value for Vb, modifying
considerably the proportions of the diagram.
Figures 103 and 104 may be conveniently condensed as
shown in Figs. 103a and 104a. The exit triangle cej of the
Fig. 104o.
previous figures is first placed with the vertex c coincident

with point a, and is then rotated about a vertical axis into
the position shown. The fact that the velocities repre-
sented by the sides of this exit triangle are now correct in
magnitude only and not in direction is of no particular
disadvantage in the study of the turbine, and this form of
diagram is very generally used. It will be termed the
polar diagram. When the entrance and exit angles of
the blades are also made equal as in this case, certain of the
lines coincide, significance of lines and
and the double
angles must be borne in mind.
Another form of diagram which is of especial value in
certain cases is shown in Fig. 105, the same nomenclature
being used as in Fig. 103. The term F„, = Viic — Vzw,
which is the total change in velocity of whirl, is here intro-
duced it is one of the factors in the expression for the work
;
done on the blade, and the directness with which it is

BLADING 99
shown on this form of diagram constitutes the special

value of the latter. Figure 106 corresponds to Fig, 104.
The effect of a variation of p = ^^ is shown by a

comparison of Figs. 103 and 105 with Figs. 104 and 106.
Fig. lOo
In the latter the value of Vb is greater and the absolute

exit velocityVz comes out with a component in the same
direction as V^- It is evident that at some intermediate
value of Vb, 1% will be perpendicular to Vb, Vzw will be zero,
90 deg., and V as shown in Figs. 107 and 108.
Fig. 107
Examination of all these figures will show that F3, < V^r,
this being the usual method of allowing for the loss of
velocity due to friction in the blade passages. The proper
coefficient to use will be the subject of a later section; it
varies from 0.70 to 0.95.
100 STEAM TURBINES
34. Work Done by Blading of a Simple -impulse Turbine.
Several methods of calculating blade work are available.

A. Since it is only the change in the value of the compo-
nent of the jet velocity parallel to the motion of the blades
which is responsible for the driving force on the blades,
the force exerted at entrance, or impulse, considered +
toward the right, is, per pound,
/^ = ^=— g g
cos «• (34)
The corresponding force at exit, or reaction, is
F,
F = ^^ = -—
F.
^ cos b- (35)
g g
(This force is + when the velocity Fs,^ is — and directed

toward the left and is — when the velocity is + and
directed toward the right; in the latter case, cos 5 is nega-
tive.) The resultant force is the vector difference of these,
and the work is this difference multiplied by the blade speed,
or
Work =E,r = — (F2. - F3.) = ~^- (36)
To calculate the work from a diagram, therefore, measure

Viw and Vzxo, take their vector difference, and multiply by
Vb/g this will give the work per pound of steam flow per
;
second.
B. An equivalent relation may be derived by substituting
for Viw and F3,„ their equivalents F2 cos a and F3 cos b, or
E,i = ^{¥2 cos « + F3 cos 5). (37)
C. Another equivalent expression may be obtained by

substituting for F2 cos a, ¥2,- cos (3 + Vb, andfor F3 cos 5,
Fsr cos 7 — Vb, giving
Ebi = ^-{V.r cos /3 + F3. cos 7). (38)

BLADING 101
D. The following method does not require the drawing of

a diagram, knowledge of the blade and nozzle angles, Vb, p
and the frictional loss only being involved.
The entrancerelative velocity V2r is diminished by
through the blade passage to the value Vsr,
friction in going
or Vsr = kbVo,, where ^"6 is a velocity coefficient for flow
through the blade passage. We have
Vb = Vsr cos y — Vs cos 8 = kbV2r cos y — Vs cos 5 (39)
Whence,
Vs cos 8 = hV.,. cos 7 - Vb, (40)
and by Eq. (37),
.
Ebi = y(F2 cos a + kbV^r COS y - V,), (41)
but
Fo cos a - Vb = Fo. COS /3,' and 7o, = ^^l^^^^-^^-^.

" (42)
COS /3
^ ^
Therefore,
Tr ^& COS 7(^2 COS a - Vb)

Ehi = -— Vo COS a — Tr
Vb -\
,
(43)
COS
= —g (F2 COS a — Ffe)( 1

^\ H
COS
^/
jS
'
(44)
^ ^
in foot-pounds per pound per second. If p is substituted

for Vb/Vo, then
F.Vcos a ^""^^^ '^^

1V1 J- (45)
E. The method does not use the trigonometrical

following
functions. Measure the velocities from the diagram.
Neglecting friction, the kinetic energy turned into work in
the blading is
R, = •^^-2/-'- (46)
102 STEAM TURBINES
If frictional loss is to be included, the above amount must
be diminished by the kinetic energy turned by friction back

into heat while crossing the blade, or by the lost work
Lost work = ^{Vl- V'lX (47)
whence
Eu = ^(F; - Vi + VI - Fp. (48)
F. The following is another analytical method. Dividing

the right-hand member of Eq. (48) into two parts,
VI — y2 F2 _ y2
Eu = ^'
+ 2g^' = ^^ + ^'^ ^^^^
2g
where Ei and Ed might be termed the ''work of the inlet
triangle "
and the '' work of the outlet triangle," respectively.
From Fig. (103), Vl = V; + Vl - 2V,V, cos a.
Whence,
= VI- VI- VI + 2V^V, cos a
Ei
2sr
Simplifying, and substituting for Vb from the equation pi =

Vb/V,
Ei = ^p, (2 cos a- Pi), (50)
or,
Ei = energy of nozzle discharge X pi (2 cos a — pi).
The coefficient is therefore termed the ''efficiency of the

triangle," and
Ei = eff. (51)
llx
The energy leaving the first triangle, corresponding to V2r,
is
2;(1
- eff.).
Multiplying this by A;- to allow for energy loss due to friction,
-=
V~
k'-il - eff.)(
/ v-\
-).
BLADING 103
From the exit triangle of Fig. 103,
K = K +K- 2^3rF. cos 7
and after substituting, in the expression for Ed simphfying,

and also substituting Vb = PaVsr, there results
Ea = -^ p. (2 COS 7 - p.), (52)
or,
Ed = energy of relative discharge from blades

X Pd (2 cos 7 - Pd).
The coefficient is termed the ''efficiency of the discharge

triangle," and
Ed = 2^ X eff. (53)
Finally, Eî = Ei + Ed.
Data on frictional loss will be given in Sec. 48.
35. Theoretical Efficiency, Impulse Blading.
To obtain the efficiency of impulse blading, any of the

above expressions for Ebi = work done may be divided by
the expression for the kinetic energy of the jet issuing from
the nozzle, or by Vl -^ 2g. Note that this gives ebi =
blade efficiency only, and does not include nozzle losses.
Thus, in the form of Eq. (36), the efficiency is
ebi = ^fV2. - ^3.). (54)

y b
In the form of Eq. (44), it is
ebi = 2p(c0S a - p)(l + k^lA.

COS
(55)
\ jS /
From this equation may be found the value of p for which
the. efficiency is a maximum, other conditions remaining
constant. Differentiating the expression with respect to
p and equating to zero,
afl + ^°Al\
COS
(cos a - 2p)
\ J
104 STEAM TURBINES
from which
^J-. (56)
For this value of p,

, X cos- «/-, ,
kh cos tX /r^ \
If 7 = /3, as for blading of symmetrical section,
(eM)_ = ^-^(1 + k,). (576)
Also, if kb = 1, as in the ideal case,
If, however, the blading section is not symmetrical, but 7

is made less than ^, called decreasing the exit angle, then
the fraction is increased and the maximum efficiency

coSjS
is raised. Other effects of decreasing the exit angle are
discussed in Sec. 44.
For other values of p than p,„„^ the blading efficiency
must be found from Eq. (55).
In the form of Eq. (48), the efficiency is
y\- Y\r^yrr-y\
(58)
^ 2
This method of calculation of the efficiency is usually the

most satisfactory one when the velocity diagrams are drawn
to scale.
It would appear that the energy remaining with the
exit steam having an absolute velocity Y z is here assumed
as a total loss. Actually, a considerable portion of it is
available in the form of entrance velocity Fi for the next

row of nozzles. in the assignment
Allowance for it is made
of heat to the stage in connection with the nozzle design,
as was indicated in Sec. 21. Data are scanty, but it appears
safe to allow from 50 to 75 per cent of the exit energy as
entrance energy for the next stage in cases where the exit
is approximately radial, and where the next stage follows
immediately. Where there is a space before the next stage,

BLADING 105
or where there is partial admission, the exit velocity is
lost, changing back to heat.

It is evident that the blading efficiency of a series of
stages may
be greater than the blading efficiency of a
single stage, due to this carry-over. Another method of
allowing for carry-over will be given in connection with
Reaction Blading, in Sec. 51.
36. Axial Thrust on Impulse Blading.
The an axial direction on a blade ring

force exerted in
isdependent on the change in value of the component of
steam velocity parallel to the axis, or, in other words, is
dependent on the difference in velocity of flow at entrance
and at exit = Fo/ — T^g/ (Fig. 103). The resulting force,
per pound of steam flowing per second, is (F2/ — F3/) -^ g.
This thrust performs no work and has to be cared for by
some form on the rotor shaft. The total
of thrust bearing
force is small, however, and
not important.
is
If the blade section is symmetrical and if friction is

neglected, the axial thrust will be zero (Figs. 103 and 105) if ;
i3
= 7 and ¥2,- = F3,, then V2/ = F3/. As friction always
exists, howêver, tending to reduce F3,, zero thrust can only
be obtained by increasing 7. It is desirable for other
reasons, however, to reduce 7 below the value of ^ or at
least not to increase it, so that axial thrust always exists
(in single flow turbines).
Another cause of a small amount of thrust called the
''suction efl"ect" will be explained in Sec. 60.
37. Simple Impulse Turbine.
As an
illustration of the foregoing principles, calculations
will be made and a diagram drawn for a simple impulse
turbine of the De Laval type (Fig. 7) with the following
data:
Initial conditions,150 lb. abs. and 50° superheat.

Exhaust pressure, 2 in. mercury.
Energy loss in nozzle, 10 per cent.
106 STEAM TURBINES
Nozzle angle, 20 deg.

kt, = Vs, -^ V2r = 0.85.
Blade speed, 1,250 ft. per second.
Symmetrical blading.
Referring to the blade speed, it has been so far assumed that
itcan have any value desired: actually, it is limited and the

value given is about at the upper limit, even for single wheels
of special material and construction.
Fig. 109.
hi = 1,222.8; cp = 1.604
ho = 895.0
qa = 327.8 B.t.u. ideally available. Allowing for

friction.
V2 = 223.8V0.90 X 327.8 = 3,842 ft. per second.
To construct the diagram (Fig. 109) lay off this value to

scale at an angle of 20 deg. with a horizontal base line.
Lay off the blade velocity = 1,250 ft. and draw the closing
side of the triangle V.r = 2,700 ft. Lay off 7 = /? = 29 deg.
7 min. and make Vsr = 0.85 = 2,700 = 2,295 ft. From the
lower extremity of this vector lay off Vb to the right: the
BLADING 107
closing line = V3 = 1,348 ft. 5 = 55 deg. 57 min., F2/ =

1,314, F3/ = 1,117, Fo. = 3,610, F3U, = 755, F„, = 3,610
+ 755 = 4,365 ft. It is evident that the blade velocity-
used, 1,250 ft., is much less than that required for maximum
efficiency. The work done per pound steam is
_ 4,365 X 1,250 ...ot,,,
This, divided by the energy of the jet, gives the efficiency of

the blading eu = 218 ^ 295.0 = 0.739. Divided by the
total heat energy available, the combined nozzle and blade
efficiency = 218 ^ 327.8 = 0.665.
38. Staging or Compounding, Object and Methods.

When steam expanded directly from a usual boiler
is
pressure down vacuum, or even merely to

to a condenser
atmospheric pressure, through a properly shaped nozzle,
the velocity generated is so large that it is practically
impossible to utilize it owing
efficiently in a simple turbine,
to the limitations on blade speed, as just illustrated. There-
fore there must be a large waste of energy in such a turbine.
Even if suitably high blade speeds could be attained to
reduce the loss at exit to a reasonable amount, the resulting
high speed of rotation would make it a difficult matter to
couple the turbine to a generator or other piece of machinery.
If the attempt is made to attain the high speed by increasing
wheel diameter and holding r.p.m. down, mechanical
difficulties arise, due to the need of small clearances, and
to the increased bulk and cost of the turbine with larger
wheels.
Two methods have been utilized to avoid these difficul-
ties both included under the general term staging. In the
:
first, called velocity staging, no attempt is made to absorb
all the kinetic energy of the steam jet in one set of blading.
The wheel is run at a reasonable speed and absorbs part of

the kinetic energy of the jet, and the exit steam, still with
high velocity, is turned parallel with its original direction
108 STEAM TURBINES
by fixed guide vanes and sent into blading on a second wheel.

This in turn abstracts part of the energy of the steam, and
the exit steam is sent to a third wheel, and so on. The
steam from the wheel has a negligibly low velocity.
final
In the other method, termed pressure staging, the total
pressure drop from boiler pressure to exhaust pressure is
broken up into several parts (two to thirty) through a series
of sets of nozzles. Following each pressure drop is a simple
turbine wheel which, with reasonable speed, can efficiently
utilize the moderate steam speed generated by the small
pressure drops.
39. Velocity Staging.
Figure 110 shows the extended form of velocity diagram

for a three-velocity stage turbine, in combination with
groups of blades at successive stages. The blade sections
are symmetrical, no allowance is made for friction, and the
value of Vb has been so chosen that there is zero exit
velocity of whirl at the final stage. Note the successive
increases in the values of /3 and 7, resulting in flatter and
thinner blades toward the final stage. Note also the
nomenclature for the various stages.
In producing this diagram, the first two triangles are
laid out as explained previously. F4 is then drawn below
the horizontal, symmetrical (in this particular case) with
F3 and equal to it (since friction is here neglected). Vb is
laid out to the left, then Vir, V-^r and finally Vb again to
the right, giving F5. The whole operation is again repeated
for the third stage. Moving blades are cross-hatched full,
fixed guide blades, dotted.

Figure 111 shows the same data laid out following the
method of Fig. 105. This form displays clearly the values
of Vr^, for the different stages; for the first it is Vou, — Vzw,
for the second, Vi^ — V5,,,, for the third, V^w — 0.
The condensed or polar form of diagram explained on
page 98 is especially well adapted for representing velocity
staging. In Fig. 113 the same data are used as in Fig. 110:
BLADING 109
Fiq.118 Fig.119
110 STEAM TURBINES
Fig. 112 shows the mtermediate stage before the exit
and superimposed on the
triangles are revolved to the right
entrance triangles.
A modification of the above condensed form of diagram
for the case where friction is allowed for, is shown in Fig.
114. Here 7„ = liV2r, Va = k'^Vs etc., the values of h
not necessarily being the same for all stages. Figure 115
shows the same data arranged like Fig. 111.
n-
BLADING 111
later stages in Fig. 121 are neither so flat nor so thin as in

Fig. 110.
As with the simple turbine, there is a most efficient value
of p, the speed ratio, for the velocity stage turbine. While
an expression for this may
be derived analytically, a
considerable number of conditions must be imposed in
order to make the solution determinate. A comparison of
Figs. 113, 114, 116, and 118, which have all been drawn
with the same scale and with zero exit velocity of whirl in
the final stage, will illustrate what may be expected. Taking
first the case of symmetrical blades and no friction (Fig.
113) it is evident that V^ = — —

Fa cos a
2^
where n is the number
of velocity stages.
If friction be included, other conditions remaining the
same, inspection of Fig. 114 will show that Vi, must be
decreased from the value used in the previous case if Fy is to
remain axial. Here kb was arbitrarily taken at 0.85 for
both blades and guides; actually, its value will increase in
the later rows.
however, friction be neglected but the exit angles
If,
of both blades and guides be diminished, other conditions

remaining the same, Fig. 116 shows that Vb must be
increased from its original value if F7 is to remain axial.
Here the exit angle of each blade (moving and stationary)
is arbitrarily made a mean between the angles made by
the absolute and relative velocities of the previous triangle.

As exit angles are generally decreased, and as friction is
always present, the effects indicated by Figs. 114 and 116
are both present in actual turbines and tend to neutralize
each other, the effect of friction generally predominating,
however. While these diagrams, drawn with zero exit ve-
locity of whirl in the final stage, do not represent exactly the
condition of highest efficiency for each case (see Sec. 35), they
closely approximate it and offer a fair comparison. We may
therefore say that Vb should be slightly less than V2 cos
a/2n for maximum efficiency.
112 STEAM TURBINES
40. Velocity Staging, Work and Efficiency.
Equation (36) for the work of a single wheel is, per pound
of steam,
E,t=^{V,^ - F3.) =-\FJ.

g 9
In the three- velocity stage turbine shown in Fig. Ill, the
total work is the sum of the works of the individual wheels,
or
E,i = ^(F2. - F3.) + -(F4. - F5.) +, etc. (59)
= — (F2.3 + F4.5 +, etc.) = -'2F. (60)
per pound of steam. The work done per pound of

total
steam may therefore be obtained by taking the sum of the
values of F,„ for each wheel and multiplying by Vb/g.
If the mean of the successive blade rings differ,
radii
each value of F„. must be multiplied by its own appropriate
value of Vb/g, since all blade rings rotate at the same
revolutions per minute.
The work done by the blading may also be computed
after the method of Eq. (48), after the diagram is drawn,
by means of the velocities.
E„ ^ ^^[V I- (Vi - VI) - {VI - VI)
-
(^l - Vl) - (Vl - Vi)
- (l"l - Vi) - F;]. (61)
^ ^
The efficiency of the blades is this quantity divided by Vl/

2g; divided by 778 {hi — h-z), which is the energy available
if
in the steam entering the nozzle, the result will be the

combined efficiency of nozzles and blading and will be less.
For a two velocity-stage wheel the blading efficiency is
VI- (VI - VI) -(VI- VI) - (F^ - VI) - VI

tbi = yTi (,oia;
41. Velocity Staging, Distribution of Work.

It is of interest to determine the distribution of work
among the successive velocity stages. In the ideal case of
BLADING 113
Figs. Ill and 113 for three stages, with symmetrical blading
and no friction, it is evident that in the first stage
In the second stage

Vt^^.r ^.r. ^Vl
Eo = ^'(47, + ,
2F5) (62)
In the third stage,

^^(^v
E, = ^{2V + I ^^
0)
_
= ^^b
These works are in the proportion 5:3:1. That is, the last
stage does only one-ninth of the total work. In a four-
stage turbine, the proportion is 7:5:3:1, the last wheel
doing only one-sixteenth of the work. While friction and
reduction of the exit angle slightly modify this, it still
remains substantially true in actual turbines, and shows
clearly the futility of the multiple-velocity stages, as
formerly used. Occasionally three stages are used, but
the two-stage arrangement, where the second wheel does
only one-fourth of the total work, is about as far as the
principle is carried in conventional designs. If lower
blade velocities must be secured other means should be
taken than velocity staging.
42. Pressure Staging.
In this method, the steam passes through a set of nozzles

and into the blading of the first wheel, then through a second
set of nozzles and into a second wheel, etc. Each portion
of the heat drop through the nozzles is so small that the
resulting steam velocity is moderate and requires only a
moderate blade velocity to utilize it efficiently. The
number of stages used depends on the total heat drop and
the blade speed to be used and varies ordinarily from 4 to
24, the higher number of stages corresponding to high steam
pressures, low back pressures and, as a general thing,
114 STEAM TURBINES
higher efficiencies. This type is sometimes called a Rateau

turbine (Figs. 9, 12 and 16).
Each stage of a pressure stage-impulse turbine really
consists of a simple impulse turbine, and all that has been
previously said of the simple impulse turbine will apply to
each stage here. One or more of the pressure stages, often
the first, Figs. 11, 13, 14 and 20, and occasionally the second
also, will have two velocity stages. If all the pressure
stages had velocity stages (Figs. 19 and 21), the combina-
tion is called a Curtis turbine, and from this, a single-
pressure stage provided with velocity stages is sometimes

called a Curtis stage.
The work done by the whole turbine is the sum of the
works of the separate pressure stages. The division of the
total heat drop into the proper portions for each stage,
equal or unequal, as may be desired, is essentially a problem
of thermodynamics and will be considered presently. This
distribution is brought about by assigning the proper areas
to the nozzles in the various stages. It will appear that it
is the sectional areas and the shapes of the passages through
the turbine with which the designer is primarily concerned,

and not so much the blade and diaphragm forms.
Velocity diagrams are generally drawn for each wheel.
If all the blade rings have the same diameter so that the
blade speed is the same throughout, then the nozzles should
be so designed that the velocity of discharge is the same for
all stages. If, in addition, the nozzle and blade angles are
all the same, the velocity diagrams will then be identical
and only one need be drawn. There are generally increases

in bladeand nozzle angles toward the exhaust end, however,
and losses of steam velocity due to friction vary with the
density, so that the velocity triangles wdll usually have
small variations.
43. Impulse Blading, Pitch and Radius of Curvature.
Since no expansion occurs in impulse blading, the passage

cross-section should be constant throughout its length. If
BLADING 115
is a circular arc, the back of the next

the face of the blade
blade should also be a circular arc concentric with the
former; the back is completed by two tangents. The pitch
and back curvature are therefore closely related, as will be
evident on comparison of the three sets of blades in Fig.
122, which have the same radius of curvature of face, but
different pitches.
There should be just
sufficient blades in a blade
ring to direct the steam flow
adequately a greater num-
;
ber will introduce unneces-

sary friction, and a smaller
number will allow the steam*
to short cut. The best pitch
has to be determined by
experiment. Pitches vary
from % in. to 1, though for
best efficiency it does not
appear advisable to go
above ^^ in. Sometimes
the pitch is made equal
to the radius of curvature
of the face. Experiments
Fig. 122.
by Briling indicate that for
symmetrical impulse blades, the following will give the most
efficient pitch:
A = ^.-
o—sm
2 /?
(63)
From Fig. 122a it will be seen that if the circular curvature

extends the full width of the face, h = 2R cos ^ = 2Pb2
sin /S cos /? = 2Ph sin 2^. Therefore,
h
P, = (64)
2 sin 2/3
For unsymmetrical blades, the entrance angle may be

taken.
116 STEAM TURBINES
The above-mentioned experiments also indicated that a

blade angle of 30 deg. was most efficient. If this be sub-
stituted in Eq. (63) for 13, then Pb = R, which justifies the

above statement.
In the usual case of velocity-staging, the last row of
blades, which has the greatest entrance angle, determines
the pitch. If this, for example, is 45 deg. then, by Eq. (64),
Pb = 0.506. The pitch of the other rows is then made the
same.
Sometimes an equivalent relation is specified, the ratio
width of the blade passage to the
of the clear transverse
radius of curvature of the face = T ^ R (Fig. 122a).
Fig. 123.
This ratio varies from 0.35 to 0.70 in practice. For velocity

staging, with two moving rows, one authority gives, for
about 500-ft. blade speed and 2,500-ft. steam speed, 0.40.
0.45, and 0.55 for the first moving, stationary, and second
moving rows, respectively.
Normally, the back of the blade is ii circular and concen-
tric with the face of the next blade ahead, completed by
two tangents (Fig. 123a). In making the sharp turn at
high velocity, however, the steam tends to crowd toward
the concave face, as shown in 1236 leaving a void in which
more or less turbulence may occur. The back of the blade
is therefore often filled out as in Fig. 123c to maintain
streamline flow.
important to note that if the edges of the blades are
It is
sharp, their number has no effect whatever on the cross-
sectional area of the steam stream. This area, for example
BLADING 117
at entrance, is equal to wheel periphery at mean height of

blades X height of blades X sin (3. The effect of edge
thickness, treated by means of a coefficient, is considered
in Sec. 46 and 62.
There no special relation between the pitch of
is
the blades and the peripheral width of the nozzle mouth, the
latter being dependent on the form of the nozzle. If the
nozzle mouth is too wide it does not direct the steam prop-
FiG. 124.
erly, and also requires a considerable excess height in the

blading immediately following, as shown in Fig. 124.
Here the blading must be high enough to receive steam
discharged from the corner x of the nozzle. To assist in
this, the nozzle is sometimes placed in a slightly oblique
position, directed inward (Figs. 11 and 19). If, however,
a conical casing is used, provision is thereby naturally

made properly to receive the steam coming slightly out-
ward from tangentially placed nozzles.
118 STEAM TURBINES
There are generally several blade pitches in the space

occupied by one nozzle mouth. One manufacturer says
that the pitch of the nozzles should be, where possible,
about three times the blade pitch of the moving row, this
in turn depending on the widths of the blades in the moving
row, and the importance of cutting down the overall length
of the turbine.
44. Impulse Blading, Angles.
1. Entrance Angle j3. —Referring to Figs. 103 to 106,
_ V2' sin a = sin a .^r>,

,
tan /3 = ^V2 =^ (65)

cos a — Vb COS a — p
If the entrance angle is made

accordance with the above,
in
the steam will enter the blading without shock. Consid-
erable importance was formerly attached to this, but it is
now considered that a moderate amount of deviation from
the shockless value causes only a negligible loss. If the
value of jS be made less than the above, the steam stream
will strike the back of the blade at entrance, exerting a
retarding effect. If the angle be made greater, it will
strike the face and tend to increase the impulse. The loss
will evidently be less in the latter case.
Taking another point of view, suppose that a turbine
has its speed increased above that defined by Eq. (65),
sometimes termed the ''synchronous speed," all other
conditions remaining the same. The backs of the blades
will run against the jet as just described. The conditions
will be as shown in Fig. 125, the full-line diagram being one
which will give no shock at entrance to the moving blades.
The actual blade entrance angle is that made by the line
ac, whereas the steam enters along ac' the component along ;
ac is as, and the normal component sc' also has a component

qc' in the line of motion but opposed to the direction of
blade motion, giving a retarding effect. In addition, the
exit velocity V3 is changed in direction to F^, reducing
Vw and with it the work done by the wheel.
BLADING 119
If the blade speed is reduced below synchronism, as in

Fig. 126, the stream will strike the face, but the component
qc' is in the direction of motion, increasing impulse. The
value of V^v also increases.
This effect of shock must be clearly distinguished from
the effect of variations in p. Maximum efficiency is
obtained when p has the correct value and any change

from this will reduce the efficiency; if the entrance blade
angle is also incorrect a further loss will result. The value
of p may be different from the proper value for maximum
^ Xs
120 STEAM TURBINES
Reference to a velocity diagram (Fig. 206) will show that

this shortening of V2 has the effect of increasing /3. There-
fore the entering steam tends to strike the back of the
blades, creating conditions similar to overspeeding, Fig.
125, and causing considerable loss. If, however, the angle
jSis made larger than indicated by the full load velocity
diagram, then at full load the action will correspond to

over-speeding, which, as stated above, causes httle reduc-
tion in efficiency. Under partial load the blade entrance
angle is approximately correct, hence the efficiency is good
then also.
practice of some manufacturers is to make the blade
The
inlettangent to the relative velocity vector at the maximum
blade height instead of at the mean height, for the following
reason: Steam issuing from a fixed nozzle having consider-
able radial height may
be assumed to have the same veloc-
ity (F2) at any Assume the nozzle angle a and
height.
also the blade entrance angle ^ constant over the entire
height. The blade velocity Vb will necessarily vary accord-
ing to the section considered, since it varies directly with
the distance from the axis of rotation. Since the longest
blade may have a height equal to two-fifths of the mean
radius, the value of Vb may be one-fifth or 20 per cent
greater at the tip, and 20 per cent less at the root, than at
the mean height. Considering conditions at the tip, it
has already been stated that the effect of overspeeding is
worse than that of underspeeding, hence in order to avoid
having the jet strike the back of the blade at entrance at
the tip, the angle there is made equal to /S in Fig. 103.
The conditions at the mean radius and at the root will
then be that of underspeeding, as in Fig. 126, and there will
be higher blade efficiency than if conditions had been
made synchronous at the mean height and overspeeding
allowed at the tip.
The above explanation presupposes the ordinary type of
blade having uniform sectional area and angles throughout
its active length. The difficulties described may be
BLADING 121
minimized by the use of warped blading, having different

angles at the tip and root. Such blades are generally drop
forged, and with this method of cross-section it is also
possible to diminish the cross-section from the root out-
ward. This decreases the stress at the root or, if desired,
allows the use of a longer blade with the same stress.
Thus, the length of a tapered blade may

be 25 to 28 per
cent of the mean blade-ring diameter. This increase in
length renders warping still more important.
As an illustration, in Fig. 127 are show^n velocity diagrams
for the tip, mean height, and root sections of a blade having
a height one-fourth the mean diameter; the speeds at tip
Fig. 127.
and root are then, respectively, 25 per cent greater

and 25 per cent less than at the mean height. The
nozzle angle is taken at 20 deg., the blade exit angle at
25 deg., and a velocity coefficient h = 0.85 for the blade
passages.
The blade-entrance angles resulting are 40, 34.5, and 31
deg., indicating the degree of warping necessary. The
angles 6 made by V3 are 126, 87, and 64 deg., which are the
angles the succeeding nozzle entrances should have at
tip, mean, and root heights if the carry-over velocity is to
be conserved.
2. Exit Angle 7. — In Fig. 128a is shown a velocity dia-
gram for the case without friction ; the blade form resulting
is also shown. In Fig. 1286 the exit angle is decreased 8
deg., resulting in a reduction of V3 from 400 to 320 ft. per
second, with a gain in diagram efficiency which may be
122 STEAM TURBINES
calculated as follows: Since there is no expansion and no
frictional loss within the blading,
T^' - VI 1,2002 - 4102

Case ('a) €6^
= = 0.883.
VI 1,2002
1,2002 315^
Case {b) eu = --
0.931.
1,200^
The change in blade forms produced by this decrease of

exit angle may be seen by comparing the blade sections in
{a) and (6), and it is evident that in the latter there is less
Fig. 128.
breadth of passage for steam flow. Therefore, to maintain

constant cross-sectional area, the height of the blading in
(6) at exit should be increased proportionately. If, in
seeking efficiency by means of reducing 7, it 7, is, made too

small, the necessary blade height becomes excessive or
impossible, and the blade channel becomes so long and so
sharply curved that increased frictional loss results.
Values of 7 in use vary from 15 to 30 deg. at high and
intermediate pressures and from 30 to 40 deg. at the low-
pressure end of the turbine, sometimes reaching 40 to 50
deg. in large turbines where maximum flow area is needed.
Two-row velocity stages may have 7 = 18 to 24 deg.
in the first moving and guide rows and 25 to 40 deg. in
BLADING 123
the second moving row, the larger values being for lower
pressures.
Axial flow re-entry wheels should obviously have sym-
metrical blades, or /3 = 7, since the steam returns through
the same blade passages in the reverse direction.
Since the breadths of blades do not vary in proportion to
the heights, marked decrease in exit angle can only be
applied to the shorter blades, as the necessary increase in
height would be prohibitive in the longer blades. As an
illustration, in Fig. 129, the long and short
blades, with the same angular divergence
Shroud
between shroud and root, provide an
increase of height of 40 per cent in case
(a),but only 12 per cent in (6).
The total angle turned through by the
steam between entrance and exit from the
blade = [180 deg. - (^ + t)] (Fig. 103) is
sometimes called the angle of deviation;
large values of this angle produce large
centrifugal action in the curved passage,
increasing the pressure on the concave side,
and causing the steam to spread in a radial
direction with consequent loss. A shroud
ring will prevent the spilling of the steam
over the ends of the blades.
Fig. 129.
N
Large angles of deviation resulting from decrease of 7
also cause a decrease of the velocity coefficient h. Refer-
ence to Eq. (55) will show that this decrease of h tends to
counteract the increasing blade efficiency m that should
result from decreasing the exit angle 7. Exact information
on the relation is not available; Fig. 137 shows Stodola's
suggestion as to how kb may be expected to vary with the
angle of deviation.
It is evident from what has been said of the practical
modifications in blade entrance and exit angles, that the
simple equal-angled blades discussed previously, are not
generally used, the entrance angle being increased, and the
124 STEAM TURBINES
exit angle decreased. The effect is therefore to twist the

blade from symmetrical position, to one similar to that
its
shown in Fig. 130, which shows a standard form used by

one manufacturer. The marked decrease or tapering in
width of the passage followed by the steam should be
noted; this is compensated by a considerable increase in
radial height at exit. For the reasons stated in a previous
paragraph such a marked reduction in passage width as is
Fig. 130.
shown in Fig. 130 is only possible in the earlier stages of

expansion, where the blade heights are reasonably small.
45. Impulse Blading, Breadth.
The breadth of blading, measured in an axial direction,

is somewhat arbitrary, depending, however, on the angle of
deviation of the steam, on the bending force exerted by the
steam, and to a greater extent, on the height. The stress
exerted on the cross-section of the blade at its root is not
greatly affected by change in cross-sectional area, since
stress due to centrifugal force predominates, and increased
:
BLADING 125
section simply means a greater load to be carried on the

new same stress. For high blade
section, resulting in the
speeds, high-grade materials must be used, capable of
enduring high stresses. The bending stress caused by steam
is very small, relatively, except in the last stage in velocity
compounding, where the blades are thin due to the large
blade angles. Blades in the second row of a two-row
wheel have occasionally broken. Vibration sometimes
occurs, causing the outer circumference of the blade ring
to wave back and forth axially and perhaps produce rub-
bing; thismust be prevented by adequate breadth of blade
and by properly tying the blades together vrith a stiff shroud
ring.
In large impulse turbines, the blades are almost never
less than î in. wide: they usually range from or 1 in. at %
the high-pressure end, through 13 4 in., Iji in. to 2 in, or
more at the low-pressure end. The breadth should never
be less than one-twelfth the height, this limitation applying
at the low-pressure end.
46. Impulse Blading, Height.
The heightof blading is dependent on the weight of

steam flowing per second, the steam volume in cubic feet
per pound, the velocity of flow, the blade angles, and the
circumferential arc of blading through which flow is taking
place. The height generally increases toward exit; this
increase may be equally above and below the mean blade-
ring diameter (Fig. 9), or it may take place partly or
entirely above the mean diameter giving a constantly
increasing mean diameter for each successive blade ring
(Figs. 14, 16 and 17).
The following symbols will be used
S = arc over which nozzle mouths extend, inches.

h„ = height of nozzles at exit.
ho = height of blades at entrance.
hz = height of blades, at exit.
126 STEAM TURBINES
rrin, rri'i, and mz are coefficients to allow for thickness of

nozzle walls at exit and blade edges at entrance
and exit, respectively (see Sec. 62).
v<2, = specific volume of steam at nozzle exit and blade
entrance (assumed same).
V3 = specific volume steam at blade exit.
of
V2, Vir, and from the diagram (Fig. 103).
Vsr are velocities
jS' = blade-entrance angle, where it is different from
the direction of V^r on the diagram.
Fig. 131.
Referring to Fig. 131, if the nozzles extend for a distance

S around the periphery, the total cross-sectional area of exit
flow from them is the sum of areas like those shown cross-
hatched in the nozzle mouths and is, allowing for nozzle-
wall thickness,
A„ =^ mJinS sin a.
Similarly, the total cross-sectional area of the corresponding

portion of the blade ring at entrance is the sum of the areas
shown cross-hatched in the blade entrances, and is A2 =
^2/12*5 sin ^'. In impulse blading there is no drop in pressure
BLADING 127
in passing through a stage, so that, neglecting frictional

reheating, the volume of steam leaving the nozzle equals
that entering the blading, or
Wv<> = m^hnVoS sin « = m^hVorS sin 13'. (66)
Therefore,
/i2 _ mnF2 sin a ^ Vj sin a /^-x
hn
~ 7^Z2727sin ^'
~ FsT^n /S''
^
the last equality being true if the coefficients vin and m^

are equal.
If ;3 = iS', the angle made by V-ir, then
V. sin « = Fo, sin ,3 = Fo/ (Fig. 103)

and
h-i = h„.
If, as is often the case, ^'> (3, then, by Eq. (67),
ll'l < hn.
That is, the height of the blading at entrance would be less
than the height of the nozzle exit. This is never advisable,

and the blade heights at entrance are always made slightly
more than the nozzle height (Fig. 124). This increase
varies from a minimum of He in. at the high-pressure end
of small turbines, up to H in. or more at the low-pressure
ends of large turbines.
Considering now^ the flow through the blade passages, the
volume entering is
Wv,
Wv2 = m.ohiVorS sin /S', whence, ho
miV^rS sin /3'
The volume leaving is

TFi^3
Wvs = msh^V^rS sin 7, whence h^ =
mzV^rS sin 7
Therefore,
hs m^VzVir sin ^'
(68)
/12 rrizV'iVzr sin 7
By combining Eq. 67 and Eq. 68 the blade exit height

may be obtained directly from the nozzle height, thus
avoiding complications which may be introduced by
128 STEAM TURBINES
arbitrary increases of blade entrance height and angle.
Thus
mnV3V2 sin a
(69)
msViVsr sin y
Reheating due to friction causes an increase of volume of

the steam passing through the blades, so that ^3 > ^'2- If
this increase be neglected, and if it also be assumed that the
Fig. 13i
edge coefficients m,,, m^ and tuz are equal, as is approxi-

mately true, then Eq. 68 reduces to
&'
hz _ V^r sin
(68a)
hi Vzr sin 7
and Eq. 69 reduces to
F2 sin a
(69a)
Vzr sin 7
Reference to Fig. 132 will show that these ratios are the
reciprocals of the ratios of the velocities of flow.
—
Miscellaneous Considerations. The minimum height of
impulse blading is from %
to î in. in small turbines, and
from \}.i to 2 in. in large turbines, being ll<2 to 2 per cent
of the mean ring diameter at the high-pressure end. Blade
BLADING 129
heights are here kept small in order to spread the nozzle

arc over as much of the periphery as possible and so cut
down the losses due to fanning by the idle blades. This is
especially important with high pressure and high super-

heat. With the same idea the diameter at the high-pres-
sure end is often made
smaller, with a steady increase in
diameter toward the low-pressure end. Starting at the
high-pressure end, the blade heights remain constant
(approximately) for several rows, until the stage is reached
at which there is full peripheral admission. From then on
blade heights are steadily increased. Figure 133 is an
I First S+cxge
Twei^f^- First Stage
Fig. 133. — Wheels and blades for 50,000-kw. turbine.
illustration of this variation in blade height and blade-ring

diameter, from a turbine similar to that in Fig. 16.
The tendency for blade heights to become excessive in
the last few stages, and the remedies used, have been dis-
cussed in Sec. 44.
47. Impulse Blading, Velocity Stage; Blade Angles, Widths,

and Heights.
Velocity stage blades, being designed to utilize high
steam velocities, must be laid out with especial care or
unnecessary loss will result due to eddying and friction.
Figure 134 shows a construction of the blading and wheel,
and Fig. 135, a nozzle block with attached guide blades.
In Fig. 136a is shown a typical velocity diagram, and the
slowing down of the steam speed is evident note that —
130 STEAM TURBINES
V5 < Vi < Vs < V2. The necessary passage cross-sec-

tional area therefore steadily increases, and while it is very
desirable for the sake of high diagram efficiency to decrease
the blade exit angles in each row, not a great deal can be
done in this direction on account of the excessive increase in
height that would be required. As examples of actual
designs, in one case where the nozzle angle is 17.5 deg.,
Fig. — Two-row
134.^ velocity-stage Fig. —
135. Reversing chamber and
wheel and blades. (General Electric guide blades. (Westinghouse Electric
Company.) and Manufacturing Company.)
the blade-exit angles on the three rows (moving, stationary

and moving) are 20, 20, and 27.5 deg., respectively; in
another case, with a nozzle angle of 18 deg., the blade-exit
angles are 22, 27, and 34 deg.
The blades are usually made broader at locations where
the steam velocity is high, as in the first of a series of velocity
stages. Proportions that represent good practice in two
velocity stage construction are: blade widths of l}i in. or
^/-i, IM and 1 in. for the first-moving, fixed, and second-
BLADING 131
moving rows respectively. Figure 1366 illustrates this

narrowing of the blades of the stationary guides and second-
moving row. The first row is generally wide with large
radius of curvature because experience has shown that this
gives increased efficiency.^ Figure 139 indicates the varia-
tion of losses that may be expected with blades of varying
widths.
The relation of the heights at exit in the successive rows
may be shown as follows; by the methods of Eqs. (66)
Fig. 136.
to (69),assuming equality of edge-thickness coefficients, and

dividing through by S, we have, for a two-row wheel hav-
ing the velocity diagram shown in Fig. 136a.
/i„F2sina2 = /13F3, sin73 = /14F4, sin 0:4 = /isFs, sin Y.î (70)
From the first two expressions,
hz = hn
F2/ _ Constant
V:.3/ F3/
From the first and third expressions,
Vof _ Constant
hi = /i„. (71)
T4/ Viy
From the first and fourth expressions.
Constant
h
The blade height at any stage is therefore equal to the
nozzle height (assuming rectangular section) times initial
'See also page 116.
132 STEAM TURBINES
velocity of flow divided by velocity of flow at point under

consideration. The ratio h^/h,, is termed the height ratio of
the stage, and its usual value, for two velocity stages, ranges
from 2.5 to 3. Some authorities state that improvement is
made by slightlychoking the flow through the three rows of

blading, which is equivalent to operating with a small
amount of reaction, but the desirable amount is not stated.
Figure 1365 illustrates a layout of the type considered.
Lead. — In impulse turbines, whenever there is partial
admission, the guide blades for the next row of moving
blades in the case of a velocity stage turbine, or the next
set of nozzles in the case of a pressure stage turbine, must be
displaced slightly around the circumference in the direction
of rotation so that the steam will be properly received by
them. While the steam is flowing through the passages
in a moving row, the passages themselves are moving and
carrying the steam along with them. This displacement
measured in inches of arc is termed lead. This action
causes the actual path of the steam passing along the rotor
to be a spiral of long pitch instead of a line parallel to the
48. Impulse Blading, Velocity Coefficient h.

Experimental investigations show that kb is determined
principally by irregularities of flow through the blade
passages which produce compressions and expansions and
eddying flow. Other factors which affect h, probably in
part due to their effect on the irregularities of flow, are
angle of deviation, relative steam velocity, breadth of
blade, radius of curvature of the face of the blade, form
and absolute dimensions of cross-section of blade passage,
thickness of blade edges and of the edges of the nozzle
partitions, and axial clearance between the nozzle exit and
the blade entrance. It is evident that a great deal of
experimental data is necessary to determine the separate
effect of each of these factors, even if this is possible, and the
presence of apparent contradictions in the results obtained
by various investigators is easily understandable.
BLADING 133
The effect of the increased angle of deviation resulting

from reduction of the exit angle 7 on the coefficient ki is
shown in Fig. 137, from Stodola, and prepared by him after
investigation of the various data available.
The variation of A:6 with V^r is shown in Fig. 138, also
from Stodola, and stated by him to be due to experiments
by the Brown Boveri Company. Dr. Stodola likewise
believes that other experimental evidence justifies him in
saying that kb increases rapidly for values of Fo,- up to the
critical, and decreases slowly for velocities beyond that.
140 110 100 80 60 40 20

1.00
^
90
.70
:
134 STEAM TURBINES
Goiidie recommends the following

0.75
£50
145
sAO
|Z5
5 15
fclO
t5
BLADING 135
The variations illustrated above are probably to be

explained in the differences in the conditions under which
the data was obtained the large turbine builders have their
:
own data which they use in their own methods

for this loss
of calculation and design and have been verified by tests
of the finished turbine.
49. Impulse Blading, Speed Ratio, Efficiency and Blade Speed.
theoretical values of blading efficiency (e^) may be

The
found from Eq. (55) for varying values of p: Eqs. (57a),
(576), and (57c) give the maximum blading efficiencies
136 STEAM TURBINES
144a for the case without friction, and in Fig. 1446 for the
case with friction.
In Fig. 145 are shown efficiency curves for a set of two-
row velocity stage-impulse blading with reduced blade-
1.00
g..90
|.80
c
C.60
30
:
BLADING 137
of Electrical Engineers.'^ The curves show the average

values of the stage efficiencies from actual practice, for single-
and two-row impulse turbines. The dotted curves indi-
cate the nozzle and blade efficiencies combined, leaving
out disc friction and fanning losses (which will be discussed
in Sec. 66). It is important to note that the value of p
formaximum stage efficiency is distinctly less than that for
the maximum blade and nozzle efficiency for the single-
row arrangement, and considerably less for the two-row
arrangement, due to the great decrease of disc and fanning
losses with decrease of speed. This indicates that for
maximum stage efficiency, which
is the thing desired, p
should be reduced below the value dictated by highest blade

efficiency only.
According to Fig. 146, for one row, the best value of p
is and for two rows, 0.22 to 0.24, close adherence
0.43 to 0.48,
to the best value being more important wdth two rows of
blading than with one row.
Values of ew taken from Figs. 141, 143, and 145 would
have to be multiplied by nozzle efficiency and would require
deductions to allow for disc friction and fanning losses
before stage efficiencies would be obtained.
The value of p finally selected for a machine is usually a
compromise between the value giving high thermodynamic
efficiency in line with the principles just discussed, but
involving a large, heavy and costly turbine, and a some-
what lower value, sacrificing something in efficiency, but
allowing a smaller, lighter and cheaper unit.
The values of p are summarized in the following table
138 STEAM TURBINES
Value of Blade Speed, Vi, for Impulse Turbines. Blade —

speeds of 400 to 850 ft. per second are in common use.
The largest velocities occur in the De Laval single-stage

impulse turbine, where 1,300 ft. per second has been success-
fully handled.
The upper limit of blade speed is naturally reached in
large turbines rather than in small ones (with the exception
of single-stage units) because it is in them that it is most
difficult to obtain sufficient area for steam flow in the last
stages. At a given speed of rotation (dictated by cycles
desired and poles on generator) the greater the allowable
blade speed, the greater may be the mean diameter of the
blade ring, the longer may be the blades, and hence the
greater the flow area through them. Also, economy is
more important in a large unit, and more expensive con-
struction is justified.
One firm uses 450 to 650

ft. per second for high-pressure,
two-row impulse wheels followed by additional single

stages, according to the uses to which the turbine is to
be put. For single-stage impulse wheels 900 ft. may be
used as an extreme on the last wheels of the largest
turbines, with blades drop forged of nickel steel, and
with special root fastening. Otherwise, 700 ft, is a better
limit,
50. Reaction Blading, Velocity Diagrams.
All of the previous diagrams have illustrated impulse tur-

bines; many of the deductions also hold for reaction blading,
which is, however, somewhat different in form. As already
stated, there is in reaction blading an increase in exit veloc-
ity from the moving blade over that at entrance, due to

expansion and consequent heat drop taking place while
the steam is passing through the blade channel. This heat
drop in the moving blades is usually, though not necessarily,
equal to the heat drop in the nozzles, or stationary blades,
as they are called, because the passages are identical in form.
In the velocity diagram, as shown in Figs. 149 and 150, the
BLADING 139
entrance and exit triangles are equal. V-2 is the velocity of
the jet as it issues from the stationary blades: For is the

relative velocity at entrance to the moving blades, V^r
is the relative velocity at exit from the moving blades
Fig. 148.
(increased as explained above), and Vz is the absolute exit

velocity. Evidently, F3, = Vo, and V3 = Vor- Also, a =
7 and ^ ^ 8. Note that F3, the absolute exit velocity,
8\\
140 STEAM TURBINES
''stage," and a pair, consisting of a stationary and a moving

ring, is called a ''double stage." We shall call such a pair a
''stage," however.^
Considering the reaction turbine as a whole, the pressure
decreases in stages, and there are generally more stages
than in the impulse turbine. If the diameters of the suc-
cessive blade rings are the same, Vb is the same for each
stage, and a single diagram like Fig. 150 will serve approxi-
mately for any of the stages. If there is a step-up in the
diameters of the blade rings, Vb will be increased, and
usually Vo will be increased in proportion. The diagram
will then be similar to Fig. 150 but enlarged in proportion
to the increase in Vb.
Since there are a large number of stages in the actual
reaction turbine, the pressure drop in each is small and is
within the critical range. All nozzles are therefore con-

verging and an examination of the blading of any reaction
turbine will prove this. were practicable to have such
If it
large drops of pressure that the critical pressure would be
passed, the forms of blade passages, moving and stationary,
would be convergent-divergent.
It is important to distinguish clearly between the terms
impulse and reaction as applied to turbine blading. Within
any one pressure stage of an impulse turbine, the pressure
at all points is the same; as there are usually holes in the
wheel on which the blades are mounted, it is impossible
entered. Sir Charles Parsons originally applied the term "impulse and
reaction turbine" to his invention, though it is commonly shortened to
"reaction turbine."
^ Attention is given in this text only to the case where fixed and moving
blade passages are practically identical, giving equal heat drops in each row
of a stage. This is not at all necessary, however, and it is possible to construct
combinations in which any proportion of the heat drop per stage takes place
in the moving row, up to the limiting case where all the drop is in the moving
blades. What are usually nozzles then become simply curved guide passages
delivering steam from one moving row to the next.This would be termed
100 per cent reaction; the other limiting case when all the heat drop takes
place in the fixed blades (nozzles) and none in the moving blades, termed
zero reaction, corresponds to the impiUse turbine. The usual form of reaction
blading then may be said to have 50 per cent reaction.
BLADING 141
for any considerable difference on the

of pressure to exist
two sides. The steam on the blading is
effect of the jet of
entirely independent of the expansive power of the steam:
a stream of any kind of particles, such as sand or shot, if
traveling at the same speed, would have the same effect.
The blade passages are uniform in cross-sectional area,
since their function is only to change the direction of the
stream.
In the reaction turbine, the stationary blades corre-
spond to the nozzles of the impulse turbine. The curved
passages in the moving blades not only absorb some of the
energy of the steam delivered to them by the nozzle ring
of fixed blading, but also generate additional kinetic energy
by allowing an expansion within themselves; some of this
energy is absorbed by the moving blades, and the steam
finally leaves the moving blades with exactly the same
velocity as it entered the fixed blades, though of course
with less pressure. To permit the expansion, the moving
blade passages are shaped like those in the stationary
blades, of nozzle form. In fact, the moving and fixed
blades are usually identical in form of cross-section, as
indicated in Fig. 148, though often of different materials
on account of the difference in stress.
51. Reaction Blading, Work and Efficiency.
The work done in each moving blade ring is (Fig. 150)
Eu = "-^.^ = ^ (^2 cos a + Vzr COS 7 — 76),
= — (2F2 cos a-Vb).

Substituting pV-^ = Vi,
E,i = ~\2p cos « - p2). (73)
To find the efficiency, let q = the heat drop per stage, then
q/2 = heat drop per row. Considering the reaction blad-
142 STEAM TURBINES
ing as a multiple nozzle, the portion e„g/2 is changed into
kinetic energy, e„ being the nozzle efficiency.
The steam leaves each row with a velocity V3 relative
to the blades in the following row, whether moving or fixed.
Not all this velocity is carried across the clearance space,
however, or, as it is expressed, the carry-over is not perfect,
but can be expressed in terms of the efficiency of carry-over
as
Carry-over = e,^^
"'
ft.-lb.i
The gain in kinetic energy in either row is then
2g
Since this is equal to the heat changed into work, we have
^ ^ = — —^ m loot pounds
and, transposing this, the heat supplied for a stage, two
rows, is
From Eqs. (73) and (74), the efficiency of reaction blading is
—g
VI
^
'^ cos a
(2p — p2)
'^ ^ o cos a
2p — p29
72
e\ 2g ) * 2
But Vl = Vl^+ X - 2^3.F, cos a. (Fig. 150)

And substituting V^ = Vzr, dividing by VI and substituting
P = V,/V,
VI
y^ — 1 + p- — 2p COS a.
Hence
2p cos a -
(75)
,(1 -h p' - 2p COS a)
' Compare this method of allowing for carry-over with that used foi
impulse turbines (p. 104).
BLADING 143
If the carry-over is perfect, then 6,^ =1.0 and eî = e„
(Eq. (75)). Under this condition, the only losses would

be those due to the nozzle action of the reaction blading,
and 66/ would be independent of the nozzle angle « and the
speed ratio p.
^
If on the other hand the carry-over is taken as zero,

then e,o= and
€bi = e„(2p COS a — p2). (75a)
This equivalent to considering a single reaction stage

is
by itself. If, in addition, the nozzle action is assumed

perfect, then e„ = 1.00 and
ew = 2p COS a — p-. (756)
Differentiating with re-

spect to p, equating to zero
and solving gives, for max- Moving
imum efficiency,
p = cos a. (76)
Substituting this value of

p in Eq. (756) gives Moving
(^w)max = COS^ a. (75c)
These values should be

compared with the values
of p and (eu)^^^ derived
in Eqs. (56) and (57c).
The blade-speed ratio p for the reaction turbine is twice the
value for the impulse turbine, but the efficiency of the stage
is the same in the two cases. It might be said that the
inherent efficiency of the ideal turbine stage, whatever its
type, is expressed by cos^ a, and the various actual efficiencies
of actual turbines result from differences in arrangements
and failures to achieve ideal conditions.
Reaction turbines have been built with p = cos a:
show the form of blading and the velocity
Figs. 151 to 153
diagrams, which should be compared with Figs. 148 to
1 True also for impulse blading.
144 STEAM TURBINES
150.' Ordinarily, however, the blade speed is lower than
this value of p would give, lowering somewhat the theoreti-
cal efficiency, but gaining other advantages.
In the first stage of a reaction turbine there must be
a greater heat drop than in the following stages, since the
BLADING 145
Friction of Steam, Nozzle Efficiency. —

The degradation of
kinetic energy of steam into heat in the blading of the
reaction turbine is not so evident as in the impulse turbine,
on account of the increase of velocity taking place in the
moving blade channels due to pressure drop. Thus, in
Fig. 149 steam entering the moving blading with the velocity
ac = Vir actually loses some of that velocity before exit,
but there is also a gain in velocity due to the fall in pressure
from c to e, which masks both this loss of kinetic energy
95
94
o 93
92
90
89
146 STEAM TURBINES
cylinder blades and spindle, and tips of spindle blades and

cylinder.
Allowance for Carry-over. — Little information is available
as to actual values of e,«, as direct experiment is practically
impossible and results derived from analyses of tests of
whole turbines have seldom been made public. Martin
mentions e,o = 0.52 as having been derived from such an
analysis in the case of a large turbine. Occasionally the
statement is made that the carry-over is practically per-
fect, that is, e,o = 1.00. This results in ebi = e,,, as already
stated.
As an example to show the effect of variation of e,o,
assume a = 15 deg., p = cos a = 0.9659, p^ = 0.9330 and
e„ = 0.90 (Fig. 154). Taking 6,« = 1.00 gives = e„ = ei,i
0.90. Taking e,„ = 0.50 and substituting in Eq. (75),
- 2 X 0.9659 X 0.9659 - 0.9330 ^

- n QO
eu 0.90^ _ Q^Q^^ ^ 0.9330 - 2 X 0.9330)
There is thus a 3 per cent drop in blading efficiency from
the ideal of 0.90 with perfect carry-over, to the case where

half the kinetic energy of carry-over is lost.
If the carry-over be taken as zero, then from Eq. (75a)
eu = 0.90(2 X 0.9659 X 0.9659 - 0.9330) = 0.84,
or a further drop of 3 per cent results.

If p is taken as 0.80 instead of the most efficient value
cos a, other quantities being the same, the values of €bi for
e,o = 1.00, 0.50, and are 0.90, 0.855, and 0.815,
respectively.
52. Reaction Blading, Speed Ratio, Efficiency and Blade Speed.
The effect on blading efficiency of varying p may be found

by means of Eq. (75). Assuming, for example, a = 15
deg., e„ = 0.90, = 0.75, and values of p varying from
e,„
0.40 to 1.20, data may be obtained for plotting curve A in

Fig. 155. Curve B is similarly obtained with e,o = 0.50.
BLADING 147
Curve C is said to have been worked out from experiment

and checked with good average practice. It includes the
effect of carry-over and is for normal 25 per cent blading,
although reasonably accurate over the greater portion
it is
of the reaction blading of a turbine when the speed ratio

p is uniform, and when the gaging is between 20 and 30
per cent. (For explanation of gaging see Sec. 53 and 54.)
90
148 STEAM TURBINES
used, about 0.75. Some large direct-connected marine
turbines formerly built have values as low as 0.30 to 0.35.
This was necessitated by the low speed required for the
propellerif a reasonably good efficiency of it was to be
obtained. Such direct-connected units are no longer used,

being replaced by geared units or electric drive.
Blade Speed Vh for Reaction Turbines. —In addition to the
general considerations discussed in connection with values
of Vb for impulse turbines (Sec. 49), it might be stated that
for all except the last low-pressure stages the blade speeds
tend to be less with reaction turbines. Since the rotor is
usually stepped, and has two or more diameters, several

blade speeds must be given. For high-pressure stages,
they vary from 100 to 130 on small machines and from 150
to 225 on large ones. At the low-pressure end where the
limiting values are reached, these were formerly 400 to 450
ft., but owing to improved design and construction, they
have been raised to from 550 to 650 ft. Even higher

values are being used with special blade material and
fastenings, and one modification has been to mount each of
the last few rows on a disc of uniform strength instead of
a drum as is more usual.
53. Reaction Blading, Form.
There is no form or geometrical construction for

definite
reaction blading. not shaped in accordance with any
It is
special theoretical considerations but has arrived at its
present form by a process of evolution. In general, it is
known that for high efficiency the direction of the blades at
entrance should be approximately parallel to the relative
discharge from the previous row, that the concave face
should be as nearly as possible a single radius, that there
must be a well-defined convergent nozzle at exit and this
must not become divergent for the longest allowable blade,
either at the tip where the spacing is greater, or at the root
where it is smaller than the mean. Nor should the nozzle
exit become divergent at any degree of gaging within the
BLADING 149
range of that particular blade section. The exit edges

should be as sharp as they can be rolled or drawn.
Manufacturers determine by tests the precise effects on
efficiency of varying slightly the shape of reaction blades,
expressing the results in the form of a series of constants,
or a curve of coefficients.
In the matter of blade exit, it has been seen that impulse
blades are designated by the exit angle. Reaction blades,
however, are classified
according to the gaging,
2a % ^^p ^^ ^^^ which is the ratio of the net
area for steam flow to the
area of the annular space
occupied by the blade ring.
Thus a 25 per cent blade
means that the net area for
flow is 25 per cent of the
annular space occupied by
V the blading. A normal 25
25% Normal Reaction Blades With

Various Gauging.s 50%, Normal Blades Gauged to 50%
Fig. 156. —
Twenty-five per cent norm Fig. 157. — Fifty
per cent noririE
reaction blades with various gagings. blades gaged to 50 per cent.
per cent blade might be gaged, for 20 per cent by closing

30 per cent by opening; the differences may be
slightly, or for
observed in Fig. 156. A 50 per cent blade might be gaged for
40 or for 60 per cent; a normal 50 per cent blade gaged to 50
per cent is shown in Fig. 157. Thus, a few different blade
sections (made in different breadths according to the length)
might be made to cover all requirements, allowing construc-
tion at low blading cost. More elaborate systems are also
used, one having a normal blading of 33, 40, 50, 66, and 75
per cent gaging, each in various breadths, giving somewhat
150 STEAM TURBINES
better blade forms and entrance angles on the average,

but at a greater expense. Various builders have their own
standards, differing slightly in section from one another,
and differing only slightly in efficiency for similar operating
conditions.
Reaction Blading Pitch. — The advisable pitch has evolved
from experience, like the blade form itself. It is seldom
less than 0.2 in. with the shortest blades and seldom more
than 1 in. for the longest of the parallel blade sections,
though pitches exceeding this are used with the longest
warped low-pressure blades. The usual pitches vary
from 0.25 to 0.7 in., being sometimes made 0.60 to 0.70
of the blade width.
With the usual forms of reaction blades, the pitch may
be easily varied by changing the thickness of the distance
pieces between them. This produces a small variation in
the area for flow, due to the thickness of the discharge edges ;
if the edge thickness were zero, change of pitch would

produce no change in the area for flow. Actually, the true
angle of the discharge stream will be changed also (see
Sec. 55).
54. Reaction Blading, Gaging.
The term gaging means the adjustment of the blade angle

after the blades are assembled, by forcibly twisting each
blade through a small angle by means of a special tool made
for the purpose. This is part of the work of the blader;
if "normal" the operation simply
the blades are gaged
corrects any inaccuracies in the assembly work. The
blader does not work to angles, but uses a gaging feeler
which gives the proper opening between the blades in a
ring, to correspond to the specified per cent of gaging.
(Gaging per cent was defined in Sec. 53.)
This possibility of changing the flow opening through the
blade ring by gaging is an advantage possessed by the
relatively light section reaction blading over the more mas-
sive impulse blading. It cheapens construction by allowing
BLADING 151
the use of a succession or group of blade rings of the same

blade section and height, the increased area required to
pass the increased vohime of steam being secured by
successive increases in the gaging in the rows. Generally,
the middle row of a group having the same height is normal
blading, earlier rows are closed, and later ones opened.
Some firms are, however, discontinuing the practice of
gaging, using sections made accurately to the angle desired.
Where the angle of twist in gaging would be considerable,
distance pieces of different form are used, thus definitely
Fig. 158. — Reaction blading in rotor. (Westinghoiise Electric and Manufac-

turmg Company.)
setting the blade at the proper angle, and allowing small

final adjustment by the usual forcible twisting. Where
especially large discharge angles are used, as at the low-
pressure end, the blades are often termed wing blades,
those with somewhat smaller angles being called ''semi-
wing" blades.
55. Reaction Blading Angles.
The entrance angles are from 5 to 15 deg. larger than the

angle made by the relative entrance velocity F20 jS' often
being about 85 deg. A moderate variation of this angle
seems to have little effect on efficiency.
As previously mentioned, the exit angles are not usually
directly stated, the per cent gaging being designated instead.
The geometric angle at exit, made by bisecting the angle
:
152 STEAM TURBINES
between the face and back of the blade at exit, is different

from and usually greater than the true exit angle shown by
direct measurements on the angle of the stream issuing from
the blading. This difference in angle, for a given blade
form, depends on the pitch of the blading. Another
method of measuring the discharge angle is by finding
sin-^ of the (gaging per cent -^ 100). Thus, with a blade
section similar to Fig. 1566 (25 per cent normal) the
probable discharge angle of the stream is about 17 deg., the
geometric outlet angle is 20 deg., and the sin"^ (25 per cent
gaging) is 14.5 deg. In like manner, the probable stream
discharge angle for the 50 per cent blading in Fig. 157 is
about 26 deg., the geometric outlet angle is 37.5 deg., and

the sin-^ gaging is 30 deg.^
56. Reaction Blading, Breadth.
The breadth of blading varies from } i to \}i in. according

to its length, discharge angle, blade speed, and the location
in the turbine (high pressure or low-pressure end). The
maximum ratio of length to breadth should not exceed
15 on low-speed turbines, and 10 or 12 on high-speed tur-
bines, preferably the former.
When a certain section is decided upon, the same shape
isreproduced in numerous widths, the following suggesting
the variation of width with height
Width, inches. 7-4 V2

Height, inches. IV2 12
The following formula for axial pitch of the rows, by

J. M. Newton, allows sufficient clearance and also allows
for the possibility of bending of the blades in an axial
direction.
p = 26-i-^m.
26-1-] in. +
-h 4
16
(77)
where h = height of blade.

1 For experimental data on actual discharge angles, see Report of Steatn
Nozzles Research Committee quoted in Sec. 22.
BLADING 153
57. Reaction Blading, Height.

As with impulse blading, the height of reaction blading
depends on the steam flow and density and on the blade-
exit angle, but there is always full peripheral admission.
In explaining the velocity diagram for reaction blading,
it was pointed out that one diagram will serve approximately
for all stages having the same blade velocity. As the steam
passes through a succession of stages, its volume steadily
increases so that, with constant flow and constant blade-exit
angles, the curve of blade lengths should be approxi-
mately that of the volume curve of the steam. It is com-
mercially impracticable to construct blades with these
minute variations in length, so that the blades are arranged
in groups or expansions, each group having the same blade
length. The middle row of the group would then have the
designed steam velocity, those previous to it a lower, and
those following it a higher steam velocity (unless gaged,
as previously explained). When these divergences from
the designed steam velocity become too great to combine
efficiently with the blade velocity, the blade height is
increased, say ^^ in., and another group is started. This
is repeated until the group height has become as great as
is mean blade-ring
considered advisable in relation to the
diameter, at which point thedrum on which the blades are
mounted is stepped up to a new diameter; this step then
has its series of groups. The blade speed is now greater,
steam speed and consequently a greater
calling for a greater
heat drop per stage. The increase of steam volume is
therefore more rapid on the second step, so that fewer
groups are necessary to bring the height to a dimension
where a second increase in diameter of the drum is advisable.
This drum likewise has its series of groups of increasing
height, each with still greater increase of steam speed on
account of greater blade speed accompanying greater steam

speed.
Minimum Height of Blading at High-pressure End of
Turbine. —Since there must be admission completely around
154 STEAM TURBINES
the blade ring, and the steam volume there is very small
first
due to high pressure, the blade heights work out very

its
small. The hmitation on this minimum blade length is due

to loss of efficiency by leakage through the clearance space.
In Fig. 159 is shown a portion of the blading. There is a
difference in pressure on the two sides of each row of blades,
both moving and fixed, and this causes a wasteful flow
through the clearance space. The necessary minimum
radial height of this clearance space depends on mechanical
Fig. 159.
features of the turbine, its length and diameter of rotor,
nature of bearings, excellence of construction, etc., and not

entire!}^ on the length of the blades. There is, it is true,
some increase in the radial clearance with long blades,
and it is greater at the low-pressure end than at the high-
pressure end, but not at all in proportion to blade height.
It is evident, therefore, that when the blade is short, the
clearance is a larger percentage of and consequently the
it,
proportion of leakage steam to the steam doing useful work

will be large. Experience has shown that the high-pressure
blading height should not be than 0.03Z), or, better,
less
0.04D, or less than î in.if reasonably good
in height,
efficiency at the high-pressure end is to be obtained. Heights
BLADING 155
as small as 0.013i) have been used in some direct-connected

slow-speed marine turbines. If the minimum allowable
height gives too great a flow, the diameter of the high-
pressure end of the rotor must be
decreased, or the blade exit angles
may be decreased by gaging, par-
tially closing the passages.
Maximum Height of Reaction
Blading. — In Sec. 44 and 46 were
discussed various considerations
affecting the height of impulse blad-
ing: many of them also apply to
reaction blading. The maximum
allowable height of parallel blading
is about one-fifth the mean blade-
ring diameter; with warped blades,
the ratio may be increased to 1:4
or 1:3.5.
In reaction turbines having a
great many stages the pressure drop
through each row is so small that a
converging nozzle is required. The
area at exit must be smaller, there-
fore,than at entrance, and there is
no need to increase the height of the
blade at exit over that at entrance.
Such turbines therefore usually have Face Back
(Concave) (Convex)
blades that are square across at their
Fig. 161. — Reaction
blad-
free ends,and the grooves for the ing,beveled and thinned tip.
(Westinghouse Electric and
blading on both rotor and casing
Manufacturing Company.)
are formed in a series of cylin-
drical steps, each step containing numerous rows of
blading.
Where higher blade speeds are desirable, justifying fewer
stages, orwhere low cost dictates fewer stages, it is espe-
cially desirableto have a uniform progression in blade
height, and in such cases blades may be mounted on a
156 STEAM TURBINES
cylindrical or conical rotor, the inside of the cylinder being

conical and the blade ends cut on a slope to correspond
and 161). The distance pieces between
(Figs. 30, 31, 35, 39,
the blades are so formed as to fill out the smooth conical
surface. This gives an increase in blade height at exit
which incidentally allows a somewhat smaller exit angle,
or lower per cent gaging, with good effect on efficiency.
As with impulse blades, when the maximum allowable
length has been reached at the low-pressure end, and fur-
ther increase in steam flow area is still required, the exit
angles are increased, or higher per cent gagings used.
This slightly increases the energy carried away by the
final exit velocity, but the relative velocity through the blade
passages is decreased, the ratio p is increased, and this gives
a small gain to partially offset the loss.
The last row or few rows of blading in reaction turbines
are sometimes made of the impulse form, because since
there no increase in velocity in the moving blades, the
is
final exit velocity can be decreased to a minimum.

The following theoretical relation between the blade
height and the mean blade-ring diameter is important.
Considering the exit velocity from any ring of blades, moving
or stationary (Fa or Vs,), this velocity, multiplied by the
sum of the cross-sectional areas of the blade passages in one
complete blade ring is equal to the total steam volume, or
Vo X Area (square feet) = Volume (cubic feet) = Wv.

irDhm sin y'-
Area =
.
:.-r~
144
Also,
y ^V, ^ IVDN
'
p 60 X 12 X p
where
D = mean diameter of blade ring, and
N = revolutions per minute.
Substituting,
Ti^ ttDN ^, irDhm sin 7
60 X 12 X p 144
BLADING 157
or
Nm sm 7
Of the quantities in the right-hand member, N is usually
assigned, W is at once calculated in the design, and p, m,
and sin 7 are constant for the greater portion of the blading,
excluding the last few rows, so that
D'h = V X constant. (79)
When the mean diameter D is constant throughout the

length of a drum, the values of h can then be easily calcu-
lated, but if the diameter is not constant, the equation must
be solved by trial as follows: First, assume a length of
blade, find D by adding it drum diameter and then
to the
calculate h from the above equation. Then correct D and
recalculate until satisfactory.
58. Reaction Blading, Axial Thrust.

While the change in velocity of flow is zero (Fig. 149)
and there is therefore no thrust on each blade ring of the
rotor due to it, there is still considerable end thrust due to
the drop in pressure from the inlet to the outlet side of each
row of moving blades, and a large end thrust due to the
steam pressure on the annular areas resulting from the
steps in rotor diameter. The low^-pressure end of the rotor
is exposed to condenser vacuum in the conventional design
(Fig. 27) This necessitates the balance or dummy pistons

.
usually provided, on reaction turbines (Fig. 28). While

one dummy piston can be provided of a proper size to bal-
ance the total thrust at one load condition, equilibrium will
not exist at other loads and hence it is usual to have a bal-
ance piston corresponding to each drum diameter of the rotor.
Mathematical analysis shows that the total thrust is
approximately balanced when the diameters of the balance
pistons are approximately equal to the average mean blade-
ring diameterson each of the drums.
In Fig. 27, which is typical, there are three dummies.
The space between the intermediate-pressure and low-pres-
158 STEAM TURBINES
sure drums is connected to the space at the face of the third

dummy by a pipe, the similar pipe for the step between the
high and intermediate drums not being shown. The outer
face of the low-pressure dummy is exposed to condenser
vacuum by the large connecting pipe. Part of the leakage
past the high-pressure dummy goes on past the intermediate
dummy, the rest is led across to the intermediate drum
blading, and there is a similar action at the low-pressure
dummy.
Various other arrangements of balance pistons are used.
Figures 35 and 37 show arrangements of two balance pis-
tons, and Figs. 30 and 38 a single balance piston. In many
of these designs the external thrust bearing is expected to
carry any unbalanced thrust occurring at various loads as
well as to provide means for adjusting axial clearance.
59. Multiple Exhaust System.

Considerable mention has been made of the fact that the
steam flow area at the low-pressure end is limited by the
allowable blade height, this limit on height arising partly
Fig. 162.
from tip divergence and partly from strength considera-

tions. One solution adopted in the past has been to divide
the steam approaching the last few stages into two parts,
leading these in opposite directions through two sets of
low-pressure blading on the same shaft, resulting in a
single double-flow turbine (Fig. 38). In large capacities
BLADING 159
the extension of this device results in a separate double-

flow unit for the low-pressure end, increasing cost, space
occupied, and complication.
Another more recent solution is the multiple exhaust
system devised by an English turbine designer, K. Bau-
mann, and represented in Figs. 162 to 167, which are taken
from a paper by the designer. Figure 162 represents the
intermediate and low-pressure portion of a turbine; the
rotor shaft is above, and steam from the high-pressure blad-
ing enters from the left, passes through the blade pas-
sages, and leaves at the right, passing downward to the
condenser.
Standard construction is employed up to the diaphragm
D, which is a special one having an annular division ring E.
The portion of the steam passing through the outer annular
F expands to the condenser pressure, exerts a driving force
on the blading and goes to the condenser.
/,
The portion steam passing through the inner annulus

of
G expands a small amount only in guides G and blading
H, the passages being shaped so as to cause the least possi-
ble loss of pressure. This results in a form of section
resembling reaction blading with 100 per cent reaction.
At diaphragm / the steam is again divided, the portion
passing through the outer annular K
expanding and doing
work on the blading A^ and entering the condenser; the inner
portion passes on through low-resistance passages in L
and M to the final diaphragm 0, where it in turn expands,
does work on blading Q, and goes to the condenser.
Some leakage occurs through the clearances between the
moving blades H
and M
and the diaphragms, but it is of
minor consequence on account of the large volume of steam
flowing.
Evidently the construction divides the low-pressure
steam into three parts, expanding each separately down to
the condenser pressure and discharging it, and thus obtain-
ing a much greater expansion without excessively long
blades. It thus increases the turbine capacity at which
.
160 STEAM TURBINES
Back pressure 1.0 in. IHcj.)
Total adiabafic heafdrop

42l2Btuper/b.
(~ Yieicjhfofsfeam in
\
^-^ per cenfofMal

Amount of worJ! done in
per cent offofal
A Efficiency ofstage
•
in per cenf
^—^Leakage /oss] '"P^^ <:«"^
Adiabafic heafdrop in
per cent of total. . .
kbsolufe pressure Ibpersq in. i.'fÔ llOS I.IOS 0491 0.^91
Fig. 164.
BLADING 161
SECTION D-D
^X
\\\l k\\\\\\\\\\\\\\\\\\\\\\\\\\\v.\Ns \\ ^^^^^^\^^'^^^^^^ vv^^ v^'^ ^-> \^^ ^
'
^'^>'-i^^'<S- ISS^^^s^ ^^
Fig. 165. — Reaction blade, Baumann type. (Westinghouse Electric and Manu-
facturing Company.)
Fig. 166. — Group of Baumann type blades. Fig. 167. —

Guide blades, Baumann
type. (Westinghouse Electric and Manu-
facturing Company.)
162 STEAM TURBINES
low-pressure steam-flow areas suffer restriction. It must

be noted, however, that this scheme also has its hmitations,
the tendency in the still larger units being to have a chok-
ing point at diaphragm D. Figures 35 to 37 show a large
turbine with Baumann type blading, built by the Westing-
house Company.
Figure 163 shows the same scheme when the exhaust is
divided into two parts only. The blading forms and
velocity diagrams are also given. In Fig. 164 are shown
the divisions of heat drop, steam flow, and work done, also
the efficiency and the leakage and leaving losses. A draw-
ing of a single moving blade is shown in Fig. 165, and a
photograph of a group of somewhat different blades in
Fig. 166. Figure 167 shows a group of Baumann type
guide blades; being the row immediately preceding the last
row of moving blades, they have only a single steam passage.
60. Clearances.
In the impulse turbine radial clearance at the blade tip

is of little consequence, since there is no tendency for the
steam to leak through this clearance (see Figs. 11, 15, 20,
and 1366). The clearance may therefore be liberal, from
0.125 to 0.25 in. —in fact, in some impulse wheels followed
by reaction wheel is smaller in diameter
stages, the impulse
than the casing and runs entirely clear of it. In recent
designs, however, the impulse wheel is well enclosed
at the idle sectors in order to decrease fanning and disc
losses.
The clearance that is important in the impulse turbine
isthe radial one at the shaft, under the diaphragm packing.
Here it varies from 0.010 to 0.040 in., according to the diam-
eter of the shaft.
The axial clearance is importance in the impulse
also of
turbine, especially at the high-pressure end. The steam
discharged from the nozzles crosses this clearance space
on the way to the moving blades, and the smaller the width
of this zone of high-velocity steam the better. With small
:
BLADING 163
clearance any tendency of the jet to spread is prevented.

The minimum vahie of this clearance is by considera-
fixed
tions of the relative expansion of the rotor and casing when
hot and cold, and when subject to variable loads. It
varies from 0.1 to 0.2 in., depending on the size of the
machine.
In the reaction turbine small radial clearance is of the
utmost importance, especially at the high-pressure end,
since there is direct leakage of steam past the blading
through the clearance space (Fig. 159). The usual mini-
mum at the high-pressure end is 0.02 in., some authorities
advising not less than 0.025. This varies for the different
stages, up to 0.100 in. at the low-pressure end of large
turbines.
The following formula by Speakman^ illustrates the
variation
Tip clearance = 0.01 in. + 0.008Z) (80)
where
D = diameter in feet.
Another formula is
Tip clearance = 0.015 in. + 0.003i) -h 0.005/?, (81)

where
h = height of blade in inches.
The axial clearance in reaction turbines is not so impor-

tant as in impulse turbines, since there is full peripheral
admission, and with the smaller steam velocity,
this,
prevents lateral spreading. Undue increase does reduce
carry-over efficiency, however.
A system termed ''end tightening" has recently been
introduced for lessening the leakage through the blade
clearances, and is illustrated in Fig. 168. The shroud rings
are turned to a thin edge which is brought close to the
raised base of the adjacent blade by thrust-bearing adjust-
1 Speakman, E. C, "The Determination of the Principal Dimensions of the
Steam Turbine with Particular Reference to Marine Work," Trans. Inst.

Eng. and Shipbuilders (Scotland), 1910.
164 STEAM TURBINES
ment. Finer clearances may be obtained, but the arrange-
ment has not yet been generally adopted. The radial
clearance is here unimportant. It would appear that the
Cylinder
raised bases of the blades form undesirable obstructions

in thepath of smooth flow of the steam.
Steam Flow through Radial Clearance of Impulse Blades.
An impulse wheel nominally has equal pressures on its
two faces, but several
phenomena make this
only approximately true.
Three causes tend to pro-
duce a flow of steam from
the clearance space on the
upstream side through
the blade passages.
A. The high- velocity
steam jet issuing from
the nozzle naturally entrains some of the steam while cross-
ing the clearance space and carries it along.
B. Unless the nozzle division plates end in knife edges,
the dead spaces behind them extend across into the space
BLADING 165
passed through by the moving blades. These latter act

somewhat like the vanes of an impeller pump, as shown in
Fig. 169, and the steam for filling the dead spaces must
come in part from the clearance space. The effect of both
A and B is intensified by the fact that the entrance height
of the blade passage is made somewhat greater than is
actually necessary to receive the steam stream. A and B
act to produce a lower pressure on the upstream side than
the stage pressure itself.
C. The leakage of steam along the shaft from stage to

stage through the diaphragm shaft packing tends to produce
a flow through the radial clearance space and into the blade
passages, due to the excess pressure on the upstream side of
the wheel.
These two effects tend to counteract, and neither could
amount to more than a few pounds per square inch, yet
over the whole disc area the total may be considerable.
Holes are usually cut in the disc at two or more points to
equalize the pressures on the two sides and so prevent
undue increase of thrust and possible deflection of the
wheel.
61. Shrouding.
Turbine blades that are left free at the outer ends suffer
from vibration which causes breakages due to fatigue, are
unsubstantial and, in the case of long thin blades, may
become spaced at the tips. It is therefore
incorrectly
customary to reinforce the blade tips by one of the follow-
ing methods:
A. A flat or formed strip is bent around the blade ring,
and tenons on the blade tips project through holes punched
in the ring; they are riveted, brazed, or welded. The
shroud is more easily fitted to blading having the same
height at entrance and exit, and the construction is some-
what stronger (Figs. 10a, 11, 134, and 170).
B. The outer ends of the blades are so formed that they
fit closely together, and a rectangular or dovetail strip is
166 STEAM TURBINES
placed in a groove turned in the outer surface and is finally

welded or brazed in place (Fig. 1366).
C. A wire or thread is drawn through holes punched in
the blades, and is brazed or lashed to the blades after the
latter are correctly spaced and gaged (Figs. 161, 166, 171,
and 172).The Westinghouse Company formerly used
a wire having a ''comma" section; after spacing, the tail of
Fig. 170. — Attaching impulse blades and shroud ring. {General Electric
Company.)
the comma was closed down on the body of the wire between
blades, preventing any movement of the latter (Fig. 173).
Shroud rings were formerly made continuous, but it has
been found preferable to make breaks in them at frequent
intervals. The ability of the thin ring to follow changes of
temperature more rapidly than can the more massive disc
or drum, which causes irregularities of expansion, will there-
fore not break the ring or loosen the blade fastenings in it
(Figs. 10a and 22).

BLADING 167
Impulse turbines usually have a shroud formed by

methods (A) or {B). Some early turbines were built
without a ring, but it was found that the steam tended to
spill over the ends of the blades. While shrouding pre-
vents this, there is good evidence that a portion of the
friction loss in blading is due to
friction against the inside of the
shroud ring. \
Some manufacturers use shroud-
ing with reaction blading (Figs. 29,
171), but it is not customary,
method (C) being more usual.
Two or even three reinforcing
168 STEAM TURBINES
62. Blade- and Guide -thickness Coefficients.
It has already been mentioned that, with impulse

blading with sharp edges, the coefficient is unity. Under
the actual conditions, however, the coefficient varies from
0.83> 0.90.
In reaction blading, the coefficient is from 0.75 to 0.85.
Since backs of reaction bhides are rounded at exit instead
Fig. 173. — Method of fastening reaction blading. (Westinghouse Electric and

Manufacturing Company.)
of being fiat like impulse blades, there some ambiguity.

is
If the angle of the back of the blade taken as the exit

is
angle, and the exit edge is sharp, then the value unity is
approximately correct. The true angle of exit however is
usually considered to be the bisector of the angles made by
the face and back, and here allowance must be made for
the thickness.
For blades of large gaging, 0.85 is justified.
63. Blade Fastenings.
Much ingenuity has been exercised in developing methods

of attachment of blades to wheels or drums which will
BLADING 169
allow the maximum possible blade speed of the rotor and

yet will be moderate in original and maintenance costs.
Tests and experience in manufacture and operation have
determined the best forms and proportions rather than
strict adherence to the usual methods of machine design.
The figures in the text cover a variety of designs.
64, Blading Materials.
Materials that have been used for reaction blades are

copper-zinc alloy, carbon steel, Osopper-clad steel, phosphor
bronze, manganese copper, and 5 per cent nickel-steel.
Special materials are pure nickel for very high superheat,
stainless steel in places where corrosion is troublesome, and
monel metal.
Reaction blades of parallel section are drawn or rolled
to form in straight bars, cut to length, the root end upset or
otherwise formed for the fastening, and the tip beveled to a
thin edge. Stationary blades are of manganese copper.
Warped or twisted form blades are forged between dies and
are usually of 5 per cent nickel-steel. Stationary warped
blades are cast of manganese-bronze and surface ground.
For impulse blades, brass, phosphor bronze, carbon steel,
and nickel steel are used. For small auxiliarj^ turbine
blades castings are sometimes used, but they are better drop
forged, the base and edges being milled. For larger tur-
bines, carbon or nickel customary, with pure nickel,
steel is
stainless steel, and monel metal also utilized as previously
described. The blades are forged and finished by milling,
or are milled from the solid bar.
Blading is subject to erosion and corrosion. The former
is a wearing action and is due to solid impurities in the steam
or to wet steam. The latter is a chemical action, and is due
largely to chemicals carried over from the boiler, to non-

homogeneity of the blade material, and to ordinary oxida-
tion at times when the turbine is not in operation.
: :
CHAPTER IV
LOSSES
The losses in nozzles and in blade and guide passages have

already been considered in Chaps. II and III. The remain-
ing losses may be considered under the heads of leakage
losses, rotational losses, steam pressure drop, radiation,
leaving loss, exhaust loss and electrical loss.
65. Leakage.
The losses which may take place through leakage are as

follows
A. High-pressure shaft gland packing.
B. Low-pressure shaft gland packing.
In addition, leakage may occur at the following points in an
impulse turbine :
C. Diaphragm labyrinth or carbon packings at shaft.

D. Diaphragm, joint at periphery with casing, and
transverse joint.
Leakage may occur at the following points in a reaction
turbine
E. Dummy pistons.
F. Blade tips.
Some illustrations of the various methods of preventing
leakage are shown in Figs. 175 to 183.
Three distinct types of packing are in use for steam tur-
bines labyrinth packing in which the steam tending to pass
:
through the leakage area is throttled many times and is in

addition made to follow a devious path; water seal, in which
a small centrifugal pump
runner on the shaft rotates in a
chamber and by establishing a high-water pressure in the
outer part of the chamber, prevents steam from passing;
170
LOSSES 171
76S432/
Fig ITS
Carbon Packing (Floating Ring Type)
\ Steam supplied
afsfarfmg \
yô^.M^
'n Cen}<rifugoil runmr^. I Carbon packmq casmgfhalf) S Stop for (5)
II I
I
Z Retaining bolfs for casing 6 Brackef
j 3 Carbon ringMSegmenfs) 7 Block for(8)
Fig.ITT 4 Garter spring for (5) 8 Supporting
dprinoj
172 STEAM TURBINES
and contact or solid packing, generally consisting of rings of

carbon or similar material fitted about the shaft and held
in a housing. The first alone cannot prevent leakage, but
can only reduce it to any desired amount depending on the
amount of labyrinth used. The second can absolutely stop
leakage, but must be supplied continuously with a small
amount of water under pressure. The third is the only one
in which there is actually physical contact; while it can be
made tight, it is subject to wear and breakage, and is not
much used in large turbines.
Shaft packing of the carbon-ring type is shown in Figs. 8,
175 and 176, and of the labyrinth type combined with water
seal in Figs. 177 and 178.
Diaphragm packings, as used on an impulse turbine, are
illustrated in Figs. 89, 90, 179 and 180.
Dummy or balance piston packings used in reaction tur-
bines are illustrated in Fig. 181. Here axial clearance is the
important one, and it can be adjusted by the locating
screws of the external thrust bearing. In Fig. 182 is shown
details of a type in which the important clearance (that
which determines leakage) is radial, cannot be adjusted,
and is dependent on the character of workmanship of the
shop. In Fig. 183, the radial clearance may be adjusted
by moving the rotor slightly endwise.
The location of the leakage at the blade tips in the reac-
tion turbine has already been discussed in Sec. 57, and is
illustrated in Figs. 159, 160, and 168.

The methods determining the magnitudes of the
of
various leakage losses will now be described.
A. The leakage through the high-pressure shaft packing
may be calculated by treating the leakage area as an orifice
with a low coefficient, say 80 per cent. Since the drop at
each contraction of the labyrinth is usually more than the
critical drop,one of the simpler formulae may be used, like
Eq. (23), Sec. 18.In the impulse turbine, steam thus leak-
ing away has come from the nozzle block and through the
first row or two rows of blades —
it performs no work in the
LOSSES 173
Shaff Dummy
Fig. 182.
174 STEAM TURBINES
blading after the first stage, unless it is led by a leak-off
pipe to a point in the turbine where the pressure is suffi-
ciently low. In the reaction turbine, steam that has leaked
through the packing has performed no work whatever.
B. At the low-pressure shaft packing, air tends to leak
inwards;it may be prevented by a water-sealed packing, or
by a steam-sealed labyrinth packing having steam above

atmospheric pressure supplied to an intermediate point in
the packing; so long as a small amount of steam leaks out-
ward, it is certain that only steam and no air can be
passing inward. This steam leakage inward has no effect
on steam-flow calculations, but may affect the vacuum
very slightly.
C. Considering now the labyrinth shaft packings at the
diaphragms in an impulse turbine, the drop in pressure from
stage to stage may or may not be greater than the critical;
in either case, from a knowledge of the clearance, the pack-
ing diameter and the steam pressures on the two sides of the
diaphragm, the steam flowing through a particular clear-
ance area may be calculated by treating it as an area with a
low coefficient, say 80 per cent. The amount of leakage is
of diminishing importance from the high-pressure to the
low-pressure end. This will be evident if we assume equal
heat drops per stage, resulting in equal velocities of flow at
the various stages. If the flow areas are also equal, there
will be equal volumes of leakage, but since the total steam
flow toward the low-pressure end is much greater in volume,
the constant leakage volume at each stage will bear a pro-
gressively smaller ratio to it. With reasonable clearances,
the leakage may amount to 3 per cent at the first
diaphragm, diminishing to perhaps 0.1 per cent at the final

stage.
Diaphragm leakage at the shaft in the impulse turbine
constitutes a deduction from the steam doing useful work.
The leakage steam, divided by the sum of the leakage steam
and the steam doing useful work, is the leakage percentage,
and this subtracted from unity gives the leakage efficiency
LOSSES 175
of the stage. This leakage efficiency approaches unity

toward the low-pressure end. The leakage steam also
tends to cause a pressure difference on the two sides of a
wheel and produce an axial thrust (see page 164).
D. Leakage past the diaphragm joints and periphery
must be considered due to defects of manufacture; to cal-
culate it requires some sort of an estimate of the leakage
areas which can only be made after a knowledge of the
workmanship. It should not be large in any case, and will
be progressively smaller toward the low-pressure end as
just explained.
E. There are one, two, or three dummy pistons in reaction
turbines, according to the type (see Figs. 27 to 30 and 35 to
40) . The pressure drop over each dummy piston is usually
larger than critical, and since the diameter is large, care-
fully designed labyrinth packings, giving a true throttling
each ring, are required to reduce leakage to a reason-
effect at
able amount. In Goudie's ''Steam Turbines" are given
several methods of determining this leakage. Martin
gives the following formula, which is said to give good
results for very low pressure drops.
(82)
W. = 0.4722
^ Vi{N + loge r)
where pi = initial pressure.
Wd = leakage in pounds per square inch per second,

r = pi -^ (final pressure beyond dummy).
Vi = specific volume of steam at pressure pi.
N = number of dummy rings.
The curves shown in Fig. 184 also offer a means of deter-
mining the dummy leakage; here leakage is assumed to vary
directly as the absolute pressure and as the clearance in
thousandths. Usually the leakage varies from 5 to 8 per
cent for the first dummy, from 4 to 7 per cent for the second,
and from 2 to 2}^ per cent for the third, the lower values
176 STEAM TURBINES
10.000
a 000
6.000
£.000
4.000
X 3.000
I" 1000
g'lOOO
^°°
I
-J 600
I* 500
I 400
<=*
300
200
100
:
LOSSES 177
calling forvery close clearances. Steam leaking through

the first dummy
does no work on the high-pressure blading,
but combines with steam passing to the intermediate-pres-
sure blading and may do work there. Steam leaking past
the second dummy may do some work in the low-pressure
blading, but steam leaking past the third dummy passes
to the condenser.
F. The leakage past the blade tips in the reaction tur-
bine is the most serious loss in this type of machine,
especially at the high-pressure end. It has been previously
explained (Sec. 57) how clearance affects leakage. Atten-
tion should further be called to the fact that the clearance
steam cuts axially across the blade tips, while the working
steam follows the blade channel. For example, with 25
per cent blading, it would appear that each square inch of
clearance area passes four times as much steam as a square
inch of blade-ring annular. Analytical investigation indi-
cates that this extreme condition does not hold, however,
but Stodola and Goudie give as a rough and ready rule, to
count 1 per cent of leakage area as 2 per cent loss of leak-
age efficiency.
The diagram shown in Fig. 185, gives the ratio
_ Working steam
~ Total steam
mean blade diam-
calculated for various blade lengths and
eters by means an equation derived by Goudie, the
of
clearance being determined by Speakman's formula (Eq.
(80) for 20-deg. blades previously given
c = 0.01 + 0.008D, D being in feet and c in inches.
66. Rotational Losses.
The important rotational losses in turbines are as follows

A. Disc friction.
B. Fanning action of idle blades.
C. Bearing friction.
i). Water-sealed gland power loss.
178 STEAM TURBINES
A and B. Disc Friction and Fanning of the Idle Blades.

Considerable power must be applied to revolve even a
smooth disc at a high speed in air, due apparently to two
causes: first, fluid friction between the air and the disc;
and second, a centrifugal pumping action of the disc, tend-

ing to cause a movement of the air along the disc radially-
outward. The latter can be much diminished, especially
at low speeds, by making the clearance on both sides of
the disc as small as possible at higher speeds this enclosing
;
is less effective. Stodola gives the following formula for the

^
horsepower lost in rotating a wheel, without blading, in air
Hp.. = O.OeOS^^^yi' (83a)
where D is the mean diameter of the blade ring in feet.

The blading on the periphery of the wheel exerts a fan-
ning action on the air and requires additional power, this
depending on the speed and the blade height. If the
blading is active all around (full peripheral admission) this
part of the loss will be negligible; if the inactive arcs are
enclosed by portions of the casing with small clearance, the
loss willbe moderate, but if the inactive portions of the
wheel rotate in free steam, the loss will be excessive.
Stodola gives for this loss
Hp.. = 0.458(|^^^y^(l - d)ch'-\ (83b)
Adding the two losses, inserting two constants, changing

to d in inches and simplifying gives
d _ .-.ch^-
Hp. = în(î^j\
f
{ .^^ +I
"^
/-,
(1 - ^)-oV
^^ 26
(84)
V 1 2,360 ^ \
where m = a constant to allow for the difference in the

medium,
= 1.0 for air and highly superheated steam in a
vacuum, steam at atmospheric
1.2 for saturated
pressure, 1.3 to 1.5 for wet steam.
^ Stodola's book should be consulted if the reader is interested in the
data from which the above equations were deduced. Owing to experimental
difficulties, results are not entirely consistent.
LOSSES 179
71 = a constant to allow for the amount of enclosing

by reduction of clearance on side of disc this is ;
highly variable but may easily be as small as

0.50 to 0.25 for small- to medium-diameter
wheels, respectively. With wheels running
free, n = 1.
Vb = blade velocity, feet per second.

d = mean diameter of blade ring in inches.
V = specific volume of steam.
d = per cent of active circumference -^ 100.
c = a constant, = 1 for a single row of blades, 1.25
for two rows and 1.8 for three rows.
h = blade height, inches.
No correction is necessary for the steam friction of the shaft

supporting an impulse wheel, on account of its small
diameter.
For a reaction turbine, the drum friction loss is so small
that it can be neglected.
A few trials with Eq. (84) will show the considerable mag-
nitude of the loss due to fanning of idle blading when the
arc of admission is small (d is then small). To obviate
this, a large arc should be used, or if this reduces the
blade height excessively, the wheel diameter can be
decreased, giving small blade-ring diameters toward the
high-pressure end of the turbine, and usually requiring
governing by throtthng rather than by cutting out nozzles.
It is usually possible onh^ on large turbines (Figs. 16, 17
and 18).
Since reaction turbines always run with full admission,
there is no correction for idle blades.
C. Bearing Friction.
Numerous formulae have been offered for the loss in
shaft bearings.There is always a perfect film of oil
between the journal and the bearing, and the work of
friction results in raising the temperature of the oil, this
heat being carried away by the water in the oil cooler through
180 STEAM TURBINES
which the oil is circulated. The surface area, velocity of

bearing surface, unit load, and oil viscosity all affect the
amount of heat generated. One formula is the following,
derived by Martin
B.t.u. loss = 0.00853 dHN, (85)
where d = diameter of journal in inches.

I = length of journal in inches.
N = r.p.m.
Note that this B.t.u. loss is entirely from useful work, and,
taking account of thermal efficiency, is roughly equivalent
to three to five times its value in heat supplied to the
turbine.
Another formula recommended by an American firm is
Hp. = i^î, (86).
where W= unit load, pounds per square inch,

Vj= speed of journal surface, feet per second,
S = Saybolt viscosity in seconds at outlet oil
temperature.
These formulae require a knowledge of the dimensions of

the bearing: for design work it is usually sufficient to
assume a shaft bearing, loss of 1 per cent of the rated load
of the turbine.
In most land-type turbines the end thrust is partially
or completely counterbalanced, the thrust bearing simply
serving to hold the rotor in a definite position and maintain
clearances. In the case of marine turbines having Kings-
bury thrust bearings, the loss in a thrust bearing may be
taken equal to the loss in one shaft bearing.
D. Water-sealed Gland-power Losses.
A water-sealed gland is virtually a small centrifugal
pump which maintains a ring of water under pressure in
the chamber surrounding the impeller. H. M. Martin
: .
LOSSES 181
gives the following formula for the loss in horsepower per

impeller
-^-im-m-mK^y (87)
where -Do = extreme outer diameter of impeller,

Di = inner diameter of rotating ring of water on one
side of impeller,
Do = inner diameter of ring on other side,
A" = r.p.m.
Such water-sealed packings are ordinarily used for packing

against vacuum or low steam
pressure. If it is attempted
to use them for high steam
pressures, the difference
between Di and Do must be
considerable and this makes
Do large, consequently the
speed of the impeller pe-
riphery is large and consider-
able power is consumed.
Hence, water sealed packings
are used for high steam pres-
Fig. 186.
sures only when preceded by
labyrinth or other form of packing with a leak-off which
will lower the pressure against the packing.
67. Steam-pressure Losses.
The throttle pressure is the pressure existing on the

boiler side of the throttle valve, the latter being
wide open.
A steam-pressure^ drop approximating 5 per cent of the
throttle absolute pressure may be assumed to take place
through the steam strainer, throttle valve, and governing
valve. In marine installations having more or less com-
plicated nianeuvering valve systems, a drop of 10 to 20 lb.
may be CQiintfid_oiL—
:
182 STEAM TURBINES

68. Radiation Losses.
Losses of heat due to radiation are comparatively small

in most eases, and may generally be neglected. The high-
pressure end is usually relatively small in diameter and is
thoroughly insulated. Turbines having a two-row impulse

wheel have an especial advantage in regard to radiation,
in that the high-temperature steam is confined in a rela-
tively small nozzle block, and after the steam has emerged
into the surrounding casing, it is so much lower in tem-
perature that the radiation losses are small.
The low-pressure casing and exhaust connection, being
filledwith steam little above atmospheric temperature,
are not insulated.
69. Leaving Loss.
The steam leaving the blading of the

kinetic energy of the
last stage at an absolute velocity V^ is a total loss, and is
generally called the leaving loss. The corresponding B.t.u.
will be
^' = (22377)
•
In any particular case this can be expressed as a percentage

of the total heat drop Qa. Assuming the value of Qa at
450 B.t.u. which will be approximately the value for large
turbines with high pressure and superheat, the percentage
loss corresponding to various leaving velocities will be as
follows
Percentage loss 1 2 3 4 5 6 7
Leaving velocity 473 671 822 949 1,061 1,163 1,255
On large machines the leaving velocity may run to 700 or

800 feet per second, corresponding to loss of about 3 per
cent at most economical load increasing to 4 per cent at maxi-
mum load. Where especial efforts are made to provide
sufficient flow area at the low-pressure stages, by double or
multiple flow, or by bleeding at the higher stages and so
LOSSES 183
decreasing the amount of steam that must be provided for,
the velocity may be reduced below 500 ft., limiting the leaving
loss to perhaps 1.5 per cent.
70. Exhaust Loss.

The steamleaving the blading in a horizontal direction
must be turned downward through 90 deg. and delivered to
SO.OOO Kiv. cf/ternator
Fig. 187. —
A 50,000-kw., three-cylinder reaction turbine, Crawford Ave.,
Chicago. Exhaust from high-pressure cylinder is reheated and returned to
intermediate pressure cylinder. Vertical condensers.
•
the condenser an appreciable distance beneath, through

the exhaust connection. To produce the velocity with
which the steam flows through this connection, past various
184 STEAM TURBINES
stays and other obstructions, requires a difference of pres-

sure, and this represents a heat drop which does no useful
work. If the exhaust passage is constricted or has an
excessive number of obstructions, the loss may be con-
siderable. Good design limits the speed of the steam in the
exhaust connection to about 300 ft. per second, corre-
sponding to about 1.8 B.t.u. drop.
To minimize the exhaust loss, the exhaust end of the
turbine must be designed to produce streamline flow and
to avoid eddies and turbulence. The 90-deg. turn has
been avoided in some recent designs by the use of two
vertical tube condensers, as shown in Fig. 187.
10,000 10,000 30,000 40,000 50,000 60,000

Capacity, Kw.
-Efficiencies of generators.
71. Electrical Loss and Summary.
The custom performance of turbine generator

of stating
units in terms of kilowatt output makes a knowledge of elec-
trical loss necessary, even though it is not strictly a turbine
loss. Brown and Drewry^ give the following rough rule for
» Brown, E. H., and W. K. Drewry, "Economy Characteristics of Stage
Feed-water Heating by Extraction," Trans. Amer. Soc. Mech. Eng., 1923.

:
LOSSES 185
generator efficiencies, which may be used in the absence

of an efficiency curve of a specific machine
Efficiency = 0.98 -
3
=JfL^ X !^-
/kilowatt rating load
(88)
^ ^
\ 17)00
Figure 188 gives the efficiency curves of several generators

and the curve of full rating efficiencies from Eq. (88).
Stodola gives the following summary of the losses in
reaction turbines, in per cent:
Table IV.— Turbine Losses

CHAPTER V
EFFICIENCY, REHEAT FACTOR, AND CONDITION LINE
The term ''efficiency" is usually deffiied as the ratio of
result to effort, but in the present subject there are so
many values which might be taken for the result, and so
many for the effort, that care must be taken to distinguish
between them. A number of efficiencies have already
been introduced and more are to follow; as the total number
will be considerable it seems desirable to give them all
comprehensive consideration at this point, and to assemble
their definitions at one location for easy reference.
72. Efficiency.
The diagram in Fig. 191 shows the flow of heat through

a three-stage turbine, the various heat quantities being
indicated by either name or symbol. In Fig. 192 a portion
of a corresponding Mollier chart is shown.
Of the total amount of heat Qi supplied to the turbine
per pound of steam, only a certain proportion Qa is avail-
able, in accordance with, the second law of thermodynamics.
Of this available heat, part is lost by leakage at the shaft
packings and by radiation, and part is passed along to the
condenser as kinetic energy and reheat and is lost, but a
large part Ei is turned into work, this being termed net
heat drop. Of this mechanical work, some is lost by
journal and gear friction, by water-seal gland work, and
by work required to drive the governor and the oil and
otherpumps; the remainder Eb is available for doing useful
—
work at the turbine coupling it is often termed ''brake
work." In the generator and other electrical equipment
there are other losses which reduce the energy delivered to
ISC)
.
EFFICIENCY, REHEAT FACTOR. AND CONDITION LINE 187
All quantities in B.t.u. per pound of steam,
tz= heat
hi
heat
supplied
heat content
by
rejected
to turbine,
steam,
of
turbine,
initia
hi = heat content of steam, final,
Qa = heat energy available for the turbine as a
whole, adiabatic drop,
mechanical energy delivered to the shaft
inside the turbine,
net heat drop for the turbine,
mechanical energy delivered by turbine
shaft at the coupling,
electrical energy delivered by the turbo-
generator at switchboard,
reheat for the turbine as a whole,
cumulative heat.
, gi2, etc. = heat energy available per stage.
, e,2, etc. = mechanical energy delivered to
shaft per stage,
= net heat drop for stage.
, qr2, etc. = reheat for a stage,
ideal thermal efficiency, based on Rankine
cycle,
internal thermal efficiency, based on energy
developed within turbine,
brake thermal efficiency, based on energy
delivered to shaft coupling,
thermal efficiency, based on
electrical
energy deUvered to switchboard,
internal efficiency, efficiency ratio
or engine efficiency,
external or brake efficiency,
stage efficiency,
nozzle efficiency,
blading or diagram efficiency,
leakage efficiency of a stage,
efficiency of carry-over of k
'
netic energy of steam.
Sfjifch-
board
Fig. 191.
188 STEAM TURBINES
the switchboard to the quantity Ee. The portion of Qi

which is rejected to the condenser is Qo.
Referring to Figs. 191 and 192, if Qa = hi — hi is the heat
ideally availableby isentropic heat drop between the initial
conditions and the final pres-
.Pz sure, the ideal thermal effi-
ciency, based on the Rankine

cycle, is
Qa
P. - =
^tr
Qi
(91)
The internal thermal effi-

effi
ciency, corresponding to the

indicated thermal efficiency
^^
^ of a reciprocating engine, is
The actual or brake thermal
Fig. 192.
efficiency, based on mechan-
ical work delivered, is
brake thermal efficiency. (93)
The actual thermal efficiency based on electrical energy

delivered is
E,.
electrical thermal efficiency. (94)
The ratio of €(» to c^, which is a relative efficiency, is
^ Ei ^Qa ^ Ei
internal efficiency. (95)
''
Q\ '
Ql Qa
This very important efficiency is also often termed the
''internal efficiency ratio" and the ''engine efficiency."

It is a measure of the closeness with which the actual
turbine approaches the ideal turbine in performance.
Similarly, the external efficiency is the ratio of Ub to e<„ or,
'"
Ê^^ Qa ^ E,
= external efficiency. (96)
Qx ' Qi Qa
EFFICIENCY, REHEAT FACTOR, AND CONDITION LINE 189
Another efficienc}^ corresponds in a way to the mechanical

efficiency of a reciprocating engine: it is the ratio of the
work delivered to the turbine coupling to the work dehvered
to the shaft or drum inside the turbine and will be called
rotational efficiency,
er = .^- = rotational efficiency. (97)
The internal efficiency e^ is also often defined in terms

of steam economies as follows; if
Sr = lb. steam per kw.-hr. for Rankine cycle = ' —'
Si = lb. steam per kw.-hr. to shaft inside turbine = V>. •
Then,
€i = internal efficiency = --= ^-. (98)
Passing now from the turbine as a whole and considering

individual stages, as, for example, the first stage in Fig. 192,
g„i will be the heat ideally available by isentropic heat

drop to the final pressure (^2) in the stage, and of this, en
is turned into useful work. The remainder g„i — en = Qri
is the reheat passed on to the next stage, and is made up of
leakage steam and heating due to steam friction. The ratio
^^
_ eo _ gal - qn ^ g^^g^ efficiency. (99)
This efficiency is also often called the hydraulic efficiency

of the turbine.
The relation of the various heat quantities in the stage is
indicated in Fig. 191 by the various streams. It is evident

that the reheat of one stage is passed along with the unavail-
able heat rejected from that stage, the two merging and
becoming the heat supplied to the next stage. The portion
qa2 is made available by the second drop in pressure, and of
this, ei2 is useful work and Qr- is rejected to the third stage,
etc. The final stage reheat g,3 passes to the condenser.
Now, it might seem that the summation of the successive
values of qa in Fig. 192 would equal Qa] owing, however,
190 STEAM TURBINES
to the fact that the hnes of constant pressure on the Mollier

chart diverge toward the right, a given pressure drop
represents a greater heat drop in these regions of greater
entropy corresponding to the later stages of a turbine.
Hence we may say that
- Sg„ = RQa, (100)
where R is than unity, termed the reheat

a constant greater
factor. This quantity will be discussed in more detail
presently it will suffice for the present to say that it varies
;
from"' unity to perhaps 1.2, with the initial and final steam
conditions, and with the excellence of design and construc-
tion of the turbine. From Fig. 192 it evidently increases
with the per cent of reheat; that is, it increases as the tur-
bine efficiency decreases.
By inspection of Figs. 191 and 192 it will be evident
that the total work delivered to the shaft, which is the sum
of the works of the individual wheels, is
Ei = 2ei = -XesQa = esXqa = eMQa, (101)
provided e^ is the same for the different stages. We may

say, therefore, that the internal efficiency, already defined in
Eq. (95), can also be expressed as
or the internal efficiency of the whole turbine is greater

than the efficiency of a. single stage in the ratio R:l. Also
R = ^^-
(103)
Another important quantity is termed cumulative heat.

If in Eq. (100) the summation Zqa stops at any stage instead
of continuing to the last one, the result is the cumulative
heat to that stage, or
Q_ = x\q^ = R{Q,yi = cumulative heat. (104)
Nozzle efficiency was discussed in Sec. 21 and blading

efficiency in Sec. 40, 49, and 52. Leakage efficiency was
defined in Sec. 65C
Efficiencies Usually Attained. —The values of ei, the

internal efficiency or efficiency ratio of turbines in actual
use, vary widely according to the size and type of the
'machines. Values approximating 85. per cent have been
claimed recently in tests of machines designed especially
for high efficiency. Values from 75 to 80 per cent should be
obtained with any of the larger units of modern type
operated under improved modern conditions, in sizes from
20,000 to 60,000 kw. With less favorable conditions as to
steam pressure, steam temperature and vacuum, they may
drop to 70 per cent. The internal efficiencies of smaller
machines are naturally less, varying from 60 to 70 per cent,
while for the small units used for driving auxiliaries, values
as low as 30 per cent may be expected.
It might be added that still higher efficiencies are
obtained with the large-size units installed in American
stations, where the tendency is to operate a few very large
units rather than a considerable number of moderate- size
ones.
73. Reheat Factor and Condition Curve.

The relation of the reheat factor to turbine design can
be best illustrated by an example.
Assume that a six-stage impulse turbine is supplied with
steam at 150 lb. abs. pressure and superheated 180°, and
is so proportioned that there is a heat drop of 68,6 B.t.u. in
each stage and that the stage efficiency is 70 per cent.

From Ellenwood's charts, the initial heat content is
1290.1 B.t.u. and initial entropy 1.676, shown by the loca-
tion of the point A in the MoUier chart Fig. 193. Sub-
tracting 68.6 from hi gives 1,221.5, the heat after adiabatic
expansion from A, the entropy still being 1.676. This
locates point Bi.
Only 70 per cent of this heat drop, shown by AD, is
turned into work, the remainder DBi = 0.30 X 68.6 =
20.58 B.t.u., being reheat and going to increase the total
heat and entropy along the constant pressure line BiCi,
192 STEAM TURBINES
the resulting heat content being 1,221.5 + 20.58 =

1,242.08 B.t.u. and the entropy becoming 1.700. This is
^
[ISO
1300
1150
1200
IISO
-g 1100
o
o
c
^ lOSO
1000
950
900
850.
—
to six, until the point Ee is reached, termed the end point.

The final pressure is 1 in. Hg; the final heat content,
1,001.98, and the final entropy, 1.864. The curve A-E, is
termed the condition curve or the expansion line, of the

turbine. The quality of the steam (degrees superheat or
per cent moisture) and the specific volume may also be
noted at each point if desired.
The calculation is advantageously arranged as shown in
Table V.
Table V. Stage Heats for Six-stage Turbine
Stage
—
194 STEAM TURBINES
The theoretically available heat is found as follows:
hi = 1,290.1 at 150 lb. and 180° superheat.

/i2 = 900.1 at 1 in. Hg and 1.676 entropy.
Qa= 390.0 available heat

/
and the reheat factor is
R 6 X 68.6
1.055.
390
This applies to a six-stage turbine. If the operation be

repeated for turbines of two, four, and twelve stages, each
time choosing (by trial) such a stage heat drop that the
final pressure resulting will be 1 in. Hg, the expansion
curves A-E^, A-Ei, and A-Eu, respectively, will result, the

individual stages not being shown for these three cases.
Table VI summarizes the results of these trials, including
Table VI. Effect of Varying Number of Stages
Number of
number of stages, and for any finite number of stages, the

value of R will be less. For a different value of e,-, the value
of R also will change, decreasing as e^ increases, toward the
value unity when e^ = 1.0. Variations in initial condi-
tions —pressure —
and superheat also affect R. Small
changes in vacuum have little effect, though R is notably
different for non-condensing conditions.
It is of great importance in steam-turbine design to be able
to predict the value of R, given the initial conditions of
the steam, the final pres-
sure and the stage effi-
ciency, without going h,
through the labor indi-

cated by Tables V and
VI Various equations
.
have been proposed, em-

pirical and otherwise, to
show what may be
expected, but a series
of curves seems most
satisfactory.
By the following ^.^-
method^ condition curves
foi? various steam con-
ditions and various stage
efficiencies are obtained on the MoUier chart graphically:
from them are obtained values of R which are plotted on
separate diagrams.
Let Fig. 194 represent the Mollier chart, AC being a
small portion of the condition curve, AB the isentropic
expansion line between the same pressures, and BC a
constant pressure line. For a given scale of h and <p, the
pressure line will have a certain slope at a given point, this
1 From an article entitled "Notes on the Comparison of Steam Turbine

Efficiencies," by E. L. Robinson, General Electric Review, July, 1926. Fig-
ures 195 to 198 are also taken from this article; the values used are based on
Goodenough's Steam Tables.
196 STEAM TURBINES
slope being dependent on the properties of steam. For a
given stage efficiency e^ the condition line will also have a
certain definite slope, and the ratio of the slopes of the two
lines at agiven point will evidently be a definite quantity
independent of the scales of the diagram, and dependent
only on the stage efficiency. Hence,
Slope of condition curve

1400
1.6 1.7 1.8 1.9 '^0

1.4 1.5
Entropy
Fig. 195. — Condition curves for various initial conditions, 1 in. Hg absolute
backpressure and 70 per cent stage efficiency.
1400
1.6 1.7 U
Entropy
Fig. 196. — Condition curve." for various initial conditions, 1 in. Hg absolute
backpressure and 80 per cent stage efficiency.
198 STEAM TURBINES
k
Figure 197 shows the reheat factors obtained by the

above method, plotted against stage efficiencies, for a
number of initial pressures and temperatures, but in all
cases for 1-in. Hg exhaust pressure. Thus, with 500
lb. per square inch initial pressure and 700°F. temperature
and a s|;age efficiency of 0.84, the reheat factor would be

1.04, giving an internal efficiency of 1.04 X 0.84 = 87.3
per cent. On the other hand, with a known internal effi-
ciency of 0.85 under the same steam conditions, one or two
trials show that the reheat factor must be 1.05, giving a
Number of Stoiqes
10 5 4 3 2 1
r>
200 STEAM TURBINES
one stage. It is roughly true that the heat recovery for a

finite number of stages is less than that for an infinite
number by a fraction equal to the reciprocal of the number
of stages of the turbine. As an example, with a reheat of
R^ = 1.05 the value of R for a ten-stage turbine would be
1.05 — 0.1 X 0.05 = 1.045. For precise work, the super-
heat may be taken into account by using Fig. 198. Both
the distribution of energy among the stages and the steam
pressure cause slight variations from these curves, but it is
not generally necessary to consider them.

74. Form of Condition Curve.
So far, it has been assumed that all stages in a turbine

have the same efficiency. This is very seldom the case, for
the following reasons:
A. Efficiencies of stages operating in the superheated
region are higher than those in the wet region.According
to Brown and Drewry:^
The internal efficiency of the turbine down to the saturation hne is 10
per cent greater than the internal efficiency of the portion in the wet region,
for straight reaction type turbines. (106)
According to Blowney and Warren:-
There is 1.15 per cent decrease in turbine stage efficiency for each 1 per
cent increase in the average moisture content in the stage. (107)

As an example, if the stage efficiency when operating
superheated were 0.85, then with 10 per cent moisture in
the steam it would be
0.85(100 - 1.15 X 10) = 75.2 per cent
B. Stage efficiencies tend to be less at high pressures due
to decreased leakage efficiencies, and due to increased rota-
tion losses in the more dense steam.
C. Stage efficiencies at the final rows of the exhaust end
tend to be lower due to greater blade-exit angles, and to
leaving loss (providing the latter is charged against the last
row of blading).
The curves A-Eo, A-E,, A-E,, and A-Ev2, in Fig. 193,
were drawn with constant stage efficiency, and it should
1 Reference note on p. 184.
2 Blowney, W. and G. B. WARREN,"The Increase in Thermal Efficiency
E.,
Due to Resuperheating in Steam Turbines," Trans. Am. Soc. Mech. Eng., 1924.
be noted that they are slightly concave upward. Higher

stage efficiency would cause the condition curve to be
steeper. Hence a straight line between A and E would
indicate lower efficiency in the high-pressure regions, with
an increase in efficiency toward the lower pressure stages.
Conversely, a greater curvature than A-E signifies high
efficiency at the upper end and low efficiency toward E.
Changes in curvature may occur: curve AJEs represents
the condition curve of an impulse turbine having a two-row
velocity stage wheel followed by eight pressure stages
of greater stage efficiency.
75. End Point.
The determination of the steam conditions at the end

of the condition curve, or, as it is called, the location of the
end-point, from a set of turbine test data, is an important
and instructive application of some of the foregoing
principles.
As an example, assume a 50,000 = kw. turbine receiving
steam at 265 lb. gage superheated 200°F. and exhausting
at 1 in. abs. On a test at 38,000 kw., which is the maximum
with the primary valve furnishing all the steam (no steam
by-passed to later stages), the steam pressure at the
primary valve (practically the entrance to the first row of
nozzles) is 226 lb. abs. (throttling having taken place from
the initial pressure). The steam rate is 10.35 lb. per
generator kilowatt-hour.
The total steam is 38,000 X 10.35 = 393,300 lb. per
hour.
From Eq. (88),
^ ^ ^
n 0-055 50,000 ^_
o OQ — _
' •
,
Generator f^
efficiency = 0.98 , X .^.^w = 0.96.
3/50^00 38,000
\ 1,000
Hence
Turbine brake kilowatts = 38,000 ^ 0.96 = 39,600.
The mechanical losses, sum of bearing friction, governor
and pump drive, gland losses and radiation, as described
:
202 STEAM TURBINES
in Chap. IV, are given approximately by Brown and

Drewry's formula (Eq. (89)). Applying this to the
present case,
Losses at rated load = 4 -^ V50, 000 ^ 1,000 = 0.6 percent,
or in kilowatts = 0.006 X 50,000 = 300 kw. Hence the
internal kilowatts delivered to the shaft within the tur-
bine = 39,600 + 300 = 39,900 kw.
Hence
Steam per internal kw.-hr. = 393,300 4- 39,900 = 9.85 lb.
and
Ei = B.t.u. per pound steam = 3413 -^ 9.85 = 346.5.
The steam is throttled from 279.7 lb. abs. 226 lb. abs.;
to
this constant heat operation is represented on Fig. 199
by the line A- A' thus locating A', the initial point of the
condition curve.
From the given conditions,
hi= 1,317.0 B.t.u.
=
/i2 879.3
Qa = 437.7
and a = 346.5 -^ 437.7 = 0.792. Also K = heat con-
tent at end point = 1317.0 - 346.5 = 970.5 B.t.u. At
1 in. abs., this corresponds to (p= 1.805, thus locating the
end point E.
Figure 199 shows the location of the initial and end
points A' and E. To draw the correct condition curve
between the two requires knowledge of type of turbine,
the heat drops per stage, and the efficiency of each stage;
in the absence of this information the following approxi-
mations may be made
A. Draw a straight connecting the points and take
line
it as the condition curve. Dividing it into two parts at
the saturation line (point C), which is at the pressure 23
lb., the efficiency ci of the portion in the superheated region
= e/qa = 160 ^ 205 = 0.780 and of the portion in the

saturated region, a = 0.815.
1350
1300
1250
1200
1150
1050-
1000-
950
900
850
1.60 1.65 1.70 1.75 1.80 1.85 .90
En+rop^
Fig. 199.
204 STEAM TURBINES
B. If the turbine is of the reaction type, by the Brown

and Drewry rule (Eq. (106)) the average internal efficiency
in the superheated region should be 10 per cent higher
than in the wet region. The single straight line of the
preceding paragraph will have to be replaced by a line
broken at the saturation line, the point Ci being selected
by trial to give the desired 10 per cent difference in effi-
ciencies of the two portions. The condition curve may then

be taken as the broken line consisting of two straight
portions; flat curves will give a closer approximation, but
they should not be tangent at Ci.
C. If an impulse turbine, follow Blowney and Warren's
rule (Eq. (107)). Assume a probable constant stage effi-
ciency for the superheated region, divide the heat drop into
blocks of about 20 B.t.u. and treat them as stages. Apply
the method illustrated by Table V, Sec. 73, at the constant
(assumed) stage efficiency down to the saturation line,
and then decreasing the stage efficiency in accordance with
the above rule for each of the wet stages. If the condition
curve so obtained does not pass through the point E

already fixed, modify the assumed constant stage efficiency
and repeat until satisfactory agreement is reached.
D. If the turbine has a two-row impulse first stage,
knowledge sufficient to locate the state point J 2 at the end
of the fo-st stage, is necessary; this may be the pressure
and quality of the steam, or the stage efficiency. The form
of the remainder of the curve J2E2 may be found by the
use of method A, B or C SiS may be appropriate.
CHAPTER VI
OUTPUT VARIATION AND GOVERNING

Turbines are started or stopped by manipulating a stop
vah'e which is tightly shut when the turbine is not operat-
ing, and is wide open when it is running; this valve is

customarily though erroneously called a throttle valve.
The amount of steam admitted to the turbine nozzles and
blading is automatically controlled by the regulating or
governing valves through which it flows after leaving the
stop valve. The position or amount of opening of these
valves is controlled by a governor which is usually of the
fly-ball centrifugal type. The control is direct in the case
of small turbines (Fig. 8) but in large machines the governor
operates the valve of a hydraulic or steam cylinder called
a servo-motor and the piston of this moves the actual
governing valves. The aim of the governing is nearly
always to maintain constant speed under varying load; in
alternating-current turbine generators the requirements
as to constancy of cycles per second and hence as to
constancy of speed, are very exacting.
76. General Considerations.
Though the inherent efficiency of the steam turbine is
nearly constant over a wide range of load, the usual methods

of governing cause large losses at low loads due to throttling.
At overloads also the efficiency falls off, principally due to
increased leaving loss. There is, therefore, for a given
turbine operating under given initial steam pressure and
temperature and given exhaust pressure, a certain load at
which the turbine requires the smallest amount of steam
per kilowatt-hour: this is termed the best load or most
efficient load (abbreviation, M. E. load) and the steam rate
205
206 STEAM TURBINES
increases at greater or smaller loads. If the turbine is to
be operated at constant load as in the case of centrifugal
pump or blower drives, or for many marine turbines, then
it is of great importance to secure the lowest possible steam
rate at that load, and economy at other loads will receive

little consideration.Turbines can seldom be operated at
constant load, however, and furthermore must have reserve
capacity. The average load must then obviously be less
than the maximum that can be carried, and often must be
less than the most efficient load. Economy at part loads
is therefore usually of great importance; economy at over-
loads, since they are usually of short duration, is of less

importance.
77. Variation of Pressures.
The turbine as a whole may be considered as a steam noz-

zle with a large pressure drop; the flow of steam through
it should then be given approximately by either Eq. (22)
or Eq. (23) of Sec. 18.

W= 0.3155Vpi ^ vi.
j_
If for Vi there be substituted its value C from the
-^ (pi)^-^
equation piVy^'^ = C, and the constants are combined, there

results
W = CApi'-^'K (108)
The flow through the turbine is therefore approximately
proportional to the inlet pressure, and tests justify this
rule. This inlet pressure is defined as the pressure before
the row of nozzles in an impulse turbine, and before
first
the firstrow of blades in a reaction turbine. Tests of

turbines also show that the inlet pressure at most efficient
load is generally less than the throttle pressure by about 5
per cent of the absolute value of the latter, this amount of
drop through the governing valve being about the minimum
consistent with reliable governing. At smaller loads the
inlet pressure -falls off with the steam flow as above explained,
due to closing of the governor valve; at greater loads it also
falls off due to additional frictional and throttling losses
OUTPUT VARIATIOX AXD GOVERXIXG 207
resulting from higher steam flow. This last loss of pres-

sure from that at the point of best economy is approxi-
mately in proportion to the squares of the steam flows at
the two loads. Thus, if at best load the throttle drop
is 15 lb. with 200,000 lb. of steam flow, then with 250,000 lb.
flow the loss will be (25 -^ 20)- X 15 = 23.5 lb.

Figure 201 shows how the pressure at any given point
such as A, 5, or C in a turbine may vary with the quantity
of steam flowing; the relation is linear as just stated.
Turbines are sometimes
provided with pressure
connections at every stage,
but more often the only
available points for meas-
uring pressures are at one
or more points near the
high-pressure end where
steam is admitted to carry
overloads, and, at from one Steam used d^ Turbine
Fig. 201.
to four points farther do\\Ti
in the expansion where steam is extracted for feed heating
purposes. The variations in the pressures at these different
points are shown in Fig. 213 for a 50,000-kw. impulse tur-
bine: the abscissae are loads, however, instead of steam
flow. This relation also is approximately linear.
78. Variation of Losses.
The losses considered in Chap. IV may be roughly divided

according to their variations with the load, as:
A Decreasing with increase of load fanning of idle blades.
. ;
B. Constant; bearing friction, pump and governor drives,

radiation, water-sealed gland work.
C. Increasing directly with increase of load steam leakage
;
at shaft packing, at stage diaphragms, or at balancing

pistons; disc friction.
D. Increasing as the square of the load; leaving and
exhaust losses.
208 STEAM TURBINES
When expressed in percentages of output, A and B decrease,

C is constant, and D increases as the load increases. In
Fig. 202, the variation of the percentages of leaving and
exhaust losses is shown by one curve, and the variation of
the sum of the other losses by
another curve. The curve of total
losses obtained by summation
shows a minimum, and this indi-
cates the load at which the turbine
will be most economical (provided
the blades and nozzles are prop-
erly designed for this load). At a
somewhat greater load the initial
cost of themachine per kilowatt will be so decreased that the
total cost ofpower produced by it per kilowatt-hour will be
a minimum, even though the leaving losses are increased.
The total amount of these losses is so small (see Table IV,
p. 185) compared with turbine output, in all but the small-
est turbines, that their variations may often be neglected.
79. Willans Line and Curve of Steam Rate.

If the steam supplied to a turbine is measured at various
loads, with constant initial steam conditions and exhaust

pressure, and the results are
plotted with total steam sup-
plied as ordinates and the kilo-
watt outputs (or fractions of
rated output) as abscissae, the
result will be a line similar to
GEL, in Fig. 203, called the
Willans line. If the steam
Output Kw. M.E.
rates are calculated from this
Fig. 203.
curve for a number of points by
dividing the ordinates by the abscissae, the result is the curve
of steam rate, or economy curve, gel, also shown in Fig. 203.
The portion GE of the Willans line between no-load and
most efficient load is usually approximately straight.
OUTPUT VARIATION AND GOVERNING 209
The intercept OG on the Y-axis is the no-load steam, or

that required to keep the turbine running at speed but
deUvering no power. The meaning of the X-axis intercept
OF will be explained presently.
The straight portion GE of the Willans hne is evidently
determined when the total steam required at any two loads
is known. The two obvious points are no-load and most
steam at these two loads
efficient load; the ratio of the total
varies with the type of turbine and
determined from tests.
is
With a straight reaction turbine it runs around 0.10.

The various methods of governing and the effect of each
on turbine economy are most easily explained by use of the
Willans line. The range from no-load to most efficient
load will be considered first.
80. Ideal Nozzle Control Governing.
Assume the ideal case of a pressure stage turbine where

there is separate control of all the nozzles at every stage,
210 STEAM TURBINES
varying arc of admissions, sufficient steam would have

to be provided to overcome them as well as to carry the load,
so that the line OB would be raised a distance OD, this
being the amount of steam necessary to overcome the
losses and drive the turbine without load. The distance
OC would then be the power which would have to be sup-
plied from an outside source to drive the turbine at operat-
ing speed, without any steam being supplied.
The steam rate curve corresponding to OB would be the
straight horizontal line ab of Fig. 205, the (constant) steam
rate being dependent merely on the slope of OB in Fig. 204.
The steam rate curve corresponding to DE would be de,
obtained by dividing the ordinates of points on DE by their
abscissae.
81. Governing by Throttling.
Assume now the other extreme, where all the steam sup-
plied to the turbine is throttled on its way to the first-stage
Fig. 20G. Fig. 207.
nozzles. As the inlet pressure decreases, the pressure drops

in each of the stages must decrease, hence the steam speeds
due to the pressure drops must also decrease. Since the
blade speed remains constant, the velocity diagrams will
change as indicated in Figs. 206 and 207 for impulse and
reaction blading respectively, hence p = Vb/Vi increases
and we have the relative velocity V^r striking the back of
the blade and causing loss of efficiency (Sec. 44). Nozzle
efficiency is also less due to overexpansion, hence both
nozzle and blading efficiency decrease as the load is
decreased. The nature of the variation of the other losses
has been discussed in Sec. 78, but for the present purpose
their total in kilowatts may be assumed constant for loads
up to the most efficient load.
The total steam flow at most efficient load will be the
same as for the nozzle-controlled turbine, since there is
then no throttling; hence the Willans line for throttle

governing will pass through E as before (Fig. 204). Since
the internal efficiency falls off at lower loads, the total
steam will not decrease so fast toward no-load as it does for
nozzle control, the Willans line being less steep, and becom-
ing EG instead of ED. The increased no-load steam OG
needs explanation; the losses have been assumed the same
as before, but the greatly lowered internal efficiency causes
a larger amount of steam to be required just to maintain
speed and overcome losses. Whether or not the line GE
remains straight depends on the manner of decrease of
efficiency; in most turbine tests it comes out straight or
with a slight upward concavity at low loads.
The intercept OF is greater than the power required from
an outside source to maintain speed with no steam supplied,
which is still OC, on account of the decreasing efficiency of
nozzles and blading at low loads.
82. Effect of Throttle -governing on Condition Curve.

Since throttling is a constant heat content operation,
the state point representing initial steam conditions on
the Mollier diagram (Fig. 208) moves horizontally to the
right when the turbine is throttle governed. If AiEi is
the condition curve for most efficient load, then AiFi and
AiGi (measured on the B.t.u. scale) are respectively the
available heat Qa and the net heat drop Ei. The Rankine
cycle efficiency Ur — AiFi -^ Qi and the internal efficiency
ti= AiGi -^ AiFi. After throttling to Ao, e,,. is very evi-
dently decreased, but the effect on €» is small; a series of
trial calculations will show that ei slowly decreases as the
initial reduced by throttling. This is verified
pressure is
by examination of the constant internal efficiency curves of

212 STEAM TURBINES
Figs. 195 and 196. As the condition curves in these figures

are drawn with constant stage efficiency, it is evident that
R — ei/es must decrease with throtthng; a few trials with
the data of Fig. 197 will verify this also.
The assumption of constant stage efficiency is of course
not justified in actual turbines, and the increased number
of stages operating in the
A,_ A2_A^
superheated region with throt-
tling causes an increase in the
average stage efficiency e,.
which may and usually does

cause an increase in e,-.
Condition curves AÊ^ and

AiEi represent the effect of
still more throttling, and
the reduction in the avail-
able heat Qa becomes very
evident.
by eight simple stages, the first wheel delivers four times

as much work as any of the others, or one-third of the work
of the turbine (see curve AJEs, Fig. 193).
condition
The Willans line would then tend to lie only one-third of
the way from ED to EG. It will not, however, pass through
E because the efficiency of a two-row wheel is less than that
of the four single-row stages it replaces, even at most effi-
cient load, and use lowers appreciably the turbine

its
efficiency. Therefore HJ (Fig. 204) will be the probable

Willans and hj (Fig. 205) the corresponding steam rate
line,
curve, poorer at most efficient load, but better at light

loads. This improved economy at light loads, combined
with decreased size and cost of the unit, is the justification
for the wide use of a 2-row first stage combined with nozzle
control (also called nozzle cut-out) governing. By increas-
ing or decreasing the first-stage diameter, its propor-
tion of the total work may be varied over wide limits.
(Compare the relative diameters of the first-stage wheels
of Figs. 14 and 20.)
With this type of governing, the power developed by
the two-row first stage increases with the number of nozzles
as more nozzles are opened, up to the point at which the
first-stage pressure is about 55 per cent of the inlet pressure.
Thereafter, as more nozzles are opened, the work per nozzle
of the first stage decreases, though the work of the stage
increases due to the additional nozzles. Finally the back
pressure rises to such a value that the total work of the
stage commences to decrease.
Complete nozzle control, even on the first stage alone, is
costly to arrange and is also unnecessary unless good
economy over a very wide range of load is desired. Use is
often made of one of the following alternatives.
A. Use throttle governing on the primary valve supplying
steam to a rather large group of nozzles and carrying loads
up to or somewhat below the most efficient load. For
greater loads, supply additional steam through individ-
ually controlled nozzles. Then for all loads above that
:
214 STEAM TURBINES
at which the primary valve is wide open, the efficiency

of the first stage is nearly constant. Due to the greater
amount steam forced through the following stages, how-
of
ever, a point is finally reached at which the internal effi-
ciency of this portion of the turbine falls off due to the
increased steam friction and leaving losses, and the decreased
nozzle and blade efficiency.
B. Arrange all the nozzles in groups which are succes-
sively gradually opened by the primary, secondary, and
tertiary valves. High first-stage efficiency is obtained at
those loads when one, two, or three of the governing valves
are wide open.
84. Overloads.
So far as we have been considering the portion of the

operating range bounded by no-load and most efficient load.
It is very desirable to be able to carry loads beyond the
most efficient for the following reasons
A. If the load is constant, the turbine can be provided
to carry this load at its most efficient point, and still have
reserve capacity for unforeseen loads, or in case of lowered
steam pressure or vacuum.
B. If the load is variable, the most efficient output of the
turbine can be arranged to equal the average load, which
obviously must be much less than the maximum.
C. The initial cost per kilowatt of the turbine and acces-
sories and the overhead are decreased by operating a
turbine at as high a load as possible.
It is now customary to rate turbines at the maximum load
they will carry: the most efficient load then from three- is
fourths to seven-eighths of the maximum, varying according

to the designs of the turbine and generators (which are more
or less dependent on the use to which the unit is to be put) ^ .
1 The limitation on output imposed by the generator must not be over-

looked. Large overloads may be safely imposed on turbines by various
means provided in the original design without much additional cost, but
the heating of the windings of a generator by heavy currents places an
:
The increased output from most efficient to maximum

may be obtained in the following ways
A. With nozzle control, either individually or in groups,
as described in A and B of Sec. 83, any desired overloads
can be carried by providing sufficient nozzles in the first
stage. The decreasing efficiency of the whole turbine at
loads beyond the most efficient load is shown by the increas-
ing steepness of the Willans line, as indicated in Fig. 203,
and is reflected in the rise in the steam-rate curve.
B. With throttling control and full peripheral admis-
sion on the first stage, the turbine could be designed for
most efficient output at throttled steam pressure, and then
the maximum load secured by raising the inlet pressure
above that for most efficient operation and thus supplying
excess pressure to the nozzles. This very convenient sys-
tem was formerly much used, but it is evident that economy
at most efficient load is sacrificed, being less than it would
be at maximum possible inlet pressure. Nor is the economy
at overloads improved, because of the rapidly lowering stage
and the rapidly increasing leaving losses. This
efficiency
arrangement also gives especially poor economy at very
light loads.
C. To secure improved economy at very light loads the
method in the preceding paragraph may be modified by
closing the primary entirely at less than half load, and
admitting steam to a small wheel ahead of the normal first
stage. This allows full or large angle admission on this
first wheel and improves the economy. A third admission
still another stage ahead may be used for the lightest loads.
This scheme is practicable only in tandem or cross-com-

pound units, where the wheels mentioned are in the high-
pressure casing. The resulting Willans line and economy
curves are shown in Fig. 209.
lute limitation on generator overloads, and the designed maximum

output must not be exceeded. If built larger than needed so as to provide
large overload capacity, then a generator habitually operates at reduced
efficiency and its initial cost is proportionally increased.
216 STEAM TURBINES
D. With throttling control and full peripheral admission
in the first stage as in the second method given, the turbine
may be designed for most efficient output with wide open
primary, throttled for partial loads. Secondary and terti-
X ^
t-a:
I.
.A Q>
through point G (the losses for small loads probably being

equal to or somewhat less than when all the steam is sup-
plied through the primary). It will curve upwârd at the
bUO
500
218 STEAM TURBINES
disregarded; as the secondary is gradually opened, the
pressure backing up on the stages between (1) and (2)

is
and decreasing their efficiencies and works. Tests show

the presence of distinct breaks in the curve when the
secondary opens; the effect is similar when (2) is wide open
and (3) starts to open. The actual Willans hne and steam
rate curves will usually resemble the examples given in
Fig. 212, w^hich are test results for a 50,000 kw. turbine.
The load ranges at which the various valves open often
overlap so much in actual turbines, however, that the
breaks are indistinct and the curve simply appears rounded.
E. Combination methods of governing involving nozzle
control up to a certain load, with by-passing at higher
loads, are also used with satisfaction.
The governing of extraction and mixed pressure turbines
involves certain additional complications which will not
be treated here.
CHAPTER VII
DESIGN
The design of a steam turbine, like that of any other

important machine, involves a judicious combination of
theory with the results of experience, governed to a great
extent by the commercial element, cost. The progress of
any particular design involves a continuous series of com-
promises between what is most efficient, what will operate
most reliably, and what will cost the least. Advancement
has naturally been gradual, because the sums of money
involved are very large, and no one builder or user is willing
to reach out very far into untried territory lest disaster
result.
As an aid to design, particularly for obtaining results
quickly, the different companies have devised methods of
calculation and have accumulated and worked up data
representing their experience and investigations. Occa-
sionally some of this information finds its way into print, but
most of it is naturally retained as valuable property of the
company compiling it. (Some valuable data have been
made public from time to time, however, by independent
experimenters and writers.) Therefore it is evident that the
methods of design outlined in the following pages are not
those of any firm building turbines, nor are the coefficients
and other data exactly those used by them, though they
are reasonably close to what might be expected.
85. Preliminaxy Data.
The problem is presented to the designer in various ways,

but ordinarily the following are specified: initial steam condi-
tions, vacuum, and capacity in horsepower or kilowatts.
The revolutions per minute may be unspecified as in marine
219
220 STEAM TURBINES
sets for electric drive or for geared drive, ormay be defi-

nitely determined by the alternating-current frequency
to be generated. Highest possible economy is usually
desired, but with a limit on the price. Of course a pur-
chaser may specify an economy which cannot be obtained
except at prohibitive cost, but since he usually must have
the turbine, he will make concessions, and competition
between manufacturers will usually produce the best propo-
sition possible with the stage of development then reached.
Numerous other specifications must be set by the designer
as he proceeds, examples being blade speed and wheel (or
drum) diameter, number of stages, type of blading and
fastening, material of blading, rotor, casing, nozzles, etc.,
systems of governing, lubrication, cooling, etc.
86. Design Problem : Pressure Stage -impulse Turbine.
Calculate the principal dimensions of the nozzles and

blading of a turbine for the following conditions:
Kilowatt delivered at shaft coupling 5,000

Revolutions per minute 2,400
Maximum blade speed, feet per second 570
Initial steam pressure, pounds absolute 150
Initial superheat, degrees F 180
Exhaust pressure, inches Hg 1
Nozzle exit angle, degrees 20

Blade exit and entrance angles equal.
All stages to have same mean blade diameter.
Nozzle control governing. for stage (1).
Turbine to be similar to Fig. 9 in general arrangement.
The problem divides itself naturally into two parts:
Part A. Heat distribution and velocity relations.

Part B. Capacity, steam flow^ and area of passages.
Part A. —From Eq. (56), for highest efficiency,
cos a 0.94 n, Ar;
In Sec. 49, however, it was shown that, including disc
friction and fanning, the value of p for maximum efficiency

DESIGN 221
should be less than the above. Adopting p = 0.42, we

have V2 = 570 -^ 0.42 = 1,357 ft. per second. Construct-
ing the entrance diagram^ of Fig. 215, scaling off V2r = 840,
and with kb = 0.925, found from Sec. 48, we have V^r =
0.925 X 840 = 777 ft. Complete the exit triangle, obtain-
ing F3 = 430.
The velocity coefficient for the nozzle is obtained from
Fig. 79, kn = 0.97: the theoretical nozzle exit velocity is
therefore 1,357 -^ =
per second, correspond-
0.97 1,400 ft.
ing to a heat drop of 39.15 B.t.u. (Eq. (5a), or Ellenwood,

Table V). The nozzle efficiency e„ = 0.97 ^ = 0.941.
Fiu. 215. Fig. 216.
The blading efficiency may be found from Eq. (55) or

Eq. (58) ; using Eq.. (58) and substituting the values from
Fig. 215,
1,3572 - 4302 - (8402 - 7772) ^ „
e. = ^ ^3Â .^ = .
0.844.
.
Table VI of Ellenwood is convenient for this calculation.

The combined efficiency of nozzles and blading is 0.941
X 0.844 = 0.794. The average stage leakage will be
assumed at 1 per cent and the loss by disc friction and
fanning at 2 per cent, giving for the stage efficiency
e, = 0.794 - 0.030 = 0.764.
The value used provisionally will be e^ = 0.76.

Next will be determined the number of stages. From
the steam chart, the heat available by adiabatic expansion
will be as follows:
hi = 1,290.1 at 150 lb. and 180° superheat.
h2 = 900 1 at 1 in. Hg and 1.676 entropy.
Qa = 390.0 = available heat.

1 A scale of 20 ft. or less to the inch should be used for this diagram.
222 STEAM TURBINES
The approximate reheat factor may be found from Fig.

197; for the initial conditions 150 lb. and 538.5°, it is approx-
imately 1.054. The number of stages is therefore
N = 390X1054
oy.lo
^ jQ ^^
Ten stages will be used, and the trial heat drop per stage
will be
390 X 1.054
= 41.1B.t.u.
10
This will cause a slight increase in V2 which may require

correction.
A table is now prepared in the same manner as was Table
V, page 193, but with the quantity 41.1 subtracted and
41.1 X (1 - .76) = 9.87 added each time for reheat, this
being performed ten times and the entropies and pressures
noted. The desired final pressure, 1 in. Hg., is slightly
passed, so a second trial with a slightly smaller heat drop
is made. (An estimate of the necessary change in the
trial made by taking one-tenth of final error in
value was
B.t.u.) With qa = 40.9 and g. = 0.24 X 40.9 = 9.81, the
results are satisfactory. The volumes are then read off,
before and after the reheat, corresponding to points Bi
and Ci, B2 and C2, etc. in Fig. 193. These are all shown
in Table VIII. The corrected value of F2 is found by
Eq. F2 = 223.8 VO.941 X 40.90 = 1,388. Figure
(24);
215 redrawn with the new value of F2 as in Fig. 216 and
is
the blade efficiency eu recalculated, but the change of

efficiency will ordinarily be so slight that it may be
neglected.
The volumes in column B are the ones theoretically
attained at exit from the nozzle, and those in column C the
ones at exit from the blades which include all the reheat.
The actual nozzle-exit volumes are somewhat greater than
the values in column B, due to nozzle friction, and using the
column B values in nozzle calculations will result in a
—
DESIGN 223
Table VIII. Stage Heats for Ten-stage Turbine
Stage
224 STEAM TURBINES
frictionand fanning has not yet taken place when the

steam leaves the blading.
The total work done per pound of steam is
^. = 10 X (40.9 - 9.81) = 310.9 B.t.u.
The internal efficiency = e^ = 310.9 -- 390.0 = 0.797.

The reheat factor i^ = 10 X 40.9 -- 390 = 1.049, or it
may be found from Eq. (103)
i^ = ei -^ e. = 0.797 ^ 0.76 = 1.049.
Allowance must still be made for the following losses in

addition to the stage losses already considered; bearing
friction 1 per cent, water-sealed gland work and high-
pressure packing loss }4 per cent, radiation }'2 per cent and
governor and pump work î per cent, or a total of 2.5
per cent.^ Deducting this from the internal efficiency
gives
6, = 0.797 - 0.025 = 0.772.
The steam per kilowatt-hour for the ideal cycle is 3,413 -^
390 = 8.75 lb. The steam per kilowatt-hour delivered at

the turbine shaft coupling is
8.76 ^ 0.772 = 11.33 lb. per hour.
Thus far, the capacity of the turbine has not entered into
the calculations, and the previous work will apply to a
turbine considerably larger or smaller than the one speci-
fied, so long as it operates under the stated conditions.
Part B. —The nozzles of the turbine must discharge
W
,,,
= —
11.33
^^
60
X 5,000 =
.. T^k"
X 60
,r ^^ ,u
15.75
r
lb. of
^
steam A
per second.
The total cross-section area of the nozzles may now be

calculated for each stage, and the results shown in
Table IX.
1 Allowance for leaving loss may be made at this point by deducting an
additional 3 per cent.
DESIGN 225
226 STEAM TURBINES
Mean diameter of blade ring = 4.53 ft. = 54.40 in.

The blade heights should therefore be not less than 0.816 in.
to 1.09 in. A nozzle height of 0.85 in. will therefore be

used for the first stage. This nozzle wiirhave a section area
of 0.85^ = 0.723 sq. in. and for the first stage there will be
required 8.61 -^ 0.723 = 11.9 nozzles. Twelve will be
used; each will then be V'8.61 -^ 12 = 0.847 in. square at
exit. The peripheral space occupied by each nozzle, or
the nozzle pitch (Fig. 131) allowing for the nozzle division
plate thickness by the use of a coefficient = 0.90, is
Pitch = 0.847 -- (sin 20° X 0.90) = 0.847 -^ (0.342 X 0.9)

= 2.75 in.
The total arc occupied by the 12 nozzles will be 12 X 2.75

= 33 in.; the complete circumference is 171 in., so that the
per cent of active periphery is 33 -^ 171 = 19.3. Figure 89
shows how, even in a divided diaphragm, the complete
circumference may be made available for nozzles, though in
some forms of construction about 2 in. blank space must be
lefton each side at the diametral joint.
The maximum number of nozzles that can be accommo-
dated is 171 ^ 2.75 = 62.2: the selection of an even number
aids in the arrangement of the divided diaphragm, so that
62 will be used in the later stages having full admission.
The calculation for the number of nozzles and the space
occupied is repeated for each stage, keeping the nozzle exit
approximately 0.847 in. square except for the small adjust-
ments necessary to provide an even whole number of
nozzles. Table X gives the results. At stage 6 a point is
reached where there is not room enough for the required
number of square section nozzles of the selected dimension,
hence the number of nozzles, thereafter, remains constant
at 62. The breadth and pitch are calculated from this
number for all the remaining stages, and the height of each
stage is increased as may be necessary to provide the
desired area. Column 11 is necessary for the calculation
of the rotation loss later on.
DESIGN 227
228 STEAM TURBINES
For stage 10, solving for sin a,
Q- AreO) yî/
" " hXSX coeff.
" 10XxT7O<~0:90 ~ ^ kq'j
or
a = 3(3.65 deg. instead of 20 deg.
This increase in nozzle angle, the nozzle pitch remaining

constant, increases the breadth of the nozzle at the exit
and at the throat. The new exit breadth = pitch X
sin a X coefficient = 2.759 X 0.597X 0.90 = 1.480 in.
The new throat breadth is 862 -=- (62 X 10.3) = 1.35

in.
These revised dimensions are given at the bottom of Table

X.
A new velocity diagram should now be drawn for stage
10, as shown in Fig. 217. The calculated blading efficiency
is 0.55, and the leaving velocity is
740 ft. per second, equivalent to

10.93 B.t.u. or 2.8 per cent of the
total available heat drop. It is
evident that the last stage does
only 0.55 -^ 0.844 = 65.5 per cent
of the work each of the others
^^^' ~^''
does, and that it does not mark
full use of the vacuum of 29 in. At part loads, however,
conditions are more favorable at this stage.
For a given blade velocity, it should be evident that the
higher the rotative speed, the smaller will be the diameter
and the more difficult it will be to obtain sufficient flow area
at the last low-pressure stages. As already stated, the
rotative speed cannot usually be changed, except to that
corresponding to the next larger even number of poles in
the generator, which may involve a prohibitive increase in
cost. If, however, for a given r.p.m., the blade speed can
be increased, the mean diameter of the blade ring will also

be increased and therefore also the allowable blade height,
thus giving considerably increased area for flow. The
steam velocity will also be increased with the higher blade
DESIGN 229
speed which also reduces the area required. Hence there

is great inducement to use high bhide speeds and thus
improve the low-pressure blading area conditions. The
only disadvantages are the increased cost of mechanical
construction and the increased leaving loss. A very small
increase of blade speed in the above problem would remove
the necessity for the large nozzle and blade angles on stage
10, and improve the stage efficiency.
—
Blading Heights. The heights of the blades at entrance
should be slightly greater than the nozzle-exit heights, as
explained on page 127. This increase will be taken at
0.1 in. In stages 1 to 5 the height of nozzle exit varies from
0.847 to 0.867: the blade entrance for these stages will be
0.95 in. high.
The blade height at exit should be greater than the
calculated entrance height on account of the increase in
volume and decrease of relative velocity due to friction,
as explainedon page 127. For stage 1, V^r = 872, V^r
= 807, h = 0.847, and the blade angles equal. Hence
h, = h^/- = 0.847 X III = 0.847 X 1.081 = 0.915 in.

V Sr 80/
This than the entrance height arbitrarily selected

is less
above, and since it is not desirable to diminish the height of
the passage through the blades, it will be kept constant for

stage 1 to 5 and made equal to 0.95 in.
In the remaining stages, however, it is advisable to increase
the heights at and the results of calculations similar
exit,
to the one above are shown in Table XI.
The mean blade height at the last stage is )2(10-4 +
11.14) = and might be termed satisfactory. The
10.77 in.,
increased height at exit is of course due to the combination

of equal angles at entrance and exit with the reduction of
relative steam velocity due to friction. If it is desired to
keep equal heights at entrance and exit, as is often done in
the last few stages, the blade exit angles must be increased,
making use of Eqs. (68) and (69).
—
230 STEAM TURBINES

Table XI. Height of Blading
Stage
number
:
DESIGN 231
packing for impulse turbines. Assuming for the present

problem a clearance of 0.015 in. and a shaft diameter of
14 in., for the diaphragm between stages 1 and 2, the
pressure being 100 lb. and the drop nearly to the critical,
the leakage will be, using three strips,
1.700 X ^^ X I^ = 2,550 lb. per hour.
The steam supplied per hour is 15.75 X 60 X 60 =

total
56,700. Dividing the above leakage by this total, the
stage leakage in per cent is 4.5. Taking two strips for
each of the remaining stages, calculating the percentages
as above and finding the average of all gives 1.4 per cent
average stage leakage for the whole turbine.
Calculation of Kilowatts Delivered to the Coupling. —A check
may now be made on the preliminary assumptions as
follows
Steam supplied to turbine = 56,700 lb. per hour. Of

this 0.5 per cent is lost by radiation;
0.5 per cent is lost by gland leakage;
1.4 per cent is lost by average stage leakage;
2.4 per cent total. 100 - 2.4 = 97.6 per cent.
56,700 X 0.976 = 55,300 lb. steaiji effective. Since the

theoretical water rate is 8.75 lb. per hour, this should
produce
55,300 ^ 8.75 = 6,323 kw.
The combined nozzle and blade efficiency is 0.794 so that

6,323 X 0.794 = 5,020 kw. is the work delivered by the
blading.
Deducting the work of disc friction,
5,020 - 205 = 4,815 kw. delivered to the shaft.
Since the last stage only does 65.5 per cent of the work done
by each of the others (Fig. 217), the work delivered to the
shaft will be reduced by the amount
(1 - 0.655) (4,815 ^ 10) = 166 kw.
232 STEAM TURBINES
Hence 4,815 — 166 = 4,749 kw. to shaft, corrected for low-
pressure stage efficiency. Allowing 2 per cent for bearing
friction, water-gland work, and pump and governor drives,
leaves
4,649 X 0.98 = 4,552 kw. delivered at the coupling.
This is much below the specified 5,000 kw., so that another
trial would have to be made, with larger allowance for
stage leakage and disc friction and including allowance for
leaving loss and reduced efficiency at last stage. The stage
and internal efficiencies would be lowered and a higher
water rate required.
In order to simplify the preceding problem somewhat, a
number of considerations were omitted:
First, no allowance was made for carry-over. If the
arrangement is like the early stages of Fig. 15, the carry-

over is probably negligible, but if like Fig. 14 or 16, then
0.50 to 0.75 is probably justified for €,„, in all stages having
full peripheral admission. In these stages the stage heat
drops may be decreased by from 0.50 to 0.75 of the heat
equivalent of the absolute exit velocity of the preceding
stage, increasing the stage efficiency.
Second, no especial provision was made for meeting
the limitation on the flow area of the last stage, and
the efficiency of this stage came out unnecessarily low.
A greater heat drop should have been provided for
the last stage, and a method of making a preliminary cal-
culation of the final stage proportions, insuring maximum
possible last stage efficiency, is given in the following
section.
Third, no allowance was made for leaving loss, which
also may be determined very closely by the above mentioned
preliminary calculation.
Fourth, blade-exit angles were not reduced. This
reduction can be considerable in the early stages, diminish-
ing gradually, until at the last stages no reduction is possible
and an increase is generally required. The result will be
an increase of average stage efficiency.
: :
DESIGN 233
Fifth, all stages were made the same diameter, whereas

decreasing the diameter of the early stages reduces the loss
due to disc friction and fanning, though it may necessitate
additional stages. The use of varying
blade rings of
diameters requires an unequal distribution of heat to the
stages, roughly in proportion to the square of the mean
diameter of the blade ring. The use of the cumulative
heat diagram described on page 253 is a convenient means
of making this distribution.
Sixth, the use of a two-row impulse wheel in place of
the early stages of the turbine in the design problem would
have many advantages (p. 213), though the steam rate
would be raised and the difficulty in the last stage some-
what aggravated.
Seventh, increase of volume due to reheat while the steam
is flowing through the nozzles and through the blade pas-
sages, was neglected in calculating the exit heights of the

nozzles and blades. The effect is negligible in this problem,
but it is given consideration in another problem (p. 243).
87. Preliminary Calculation of Final Stages.
It is often very desirable to make a preliminary estimate

of the proportions of the final stages in order to determine
the possibility of a proposed turbine design, and if the
design is possible, of the efficiency and heat assignments
for the final stages before proceeding with the design of the
preceding stages.
Suppose it is proposed to design a turbine of the following
proportions
Revolutions per minute 3 600

,
Mean diameter of last blade ring, inches 32

Maximum allowable blade height, inches 6
Estimated steam flow, pounds per second 10
Final pressure, inches Hg 1
Estimated final entropy (end point) 1 .75
Leaving loss, B.t.u 8
The possibility of the design can be investigated as follows

234 STEAM TURBINES
The projected area of the last blade ring available for

steam flow, allowing for thickness of blade edges by use of
the blade thickness coefficient m = 0.9, is
Circumference X blade height X m = 32XxX6X0.9

= 543 sq. in.
The leaving velocity corresponding to 8 B.t.u. is 633 ft. per

second = V3. When the leaving velocity F3 is parallel
to the axis, the maximum possible steam flow, is taking
place, hence
Maximum steam flow = 633 X (543 ^ 144) = 2,385 cu.
ft. per second. Since the specific volume at 1 in. Hg. and
(f
= 1.75 = 560 cu. ft., the above flow corresponds to
2,385 -^ 560 = 4.26 lb. per second.The proposed design
requiring 10 lb. per second is therefore impossible.
The design few stages for the maximum pos-
of the last
sible flow, 4.26 lb. per second, will be carried on. The
velocity diagram in Fig. 218 is started with F3 axial and
Fft = (32 ^ 12) X TT X 3,600 ^ 60 = 503 ft. The clos-
ing line of the exit triangle is Vsr = 809 ft. For this
relative velocity, kb = 0.91 from Sec. 48 and For = 809
^ 0.91 = 890. Also sin 7 = 633 4- 809 = 0.783 and 7 =
51.5°.
The nozzle-exit height will be taken 0.2 in. less than the
blade height, or 5.8 in. The blade angle at entrance will be
calculated as if the height were 5.8 in. This entrance
DESIGN 235
angle 13 may be derived from Eq. (G8), reheat in the blade

passage being neglected, and blade thickness coefficients
being assumed equal.
/?3 V.r sin /5 6 890 X sin
or
/i2 Vzr sin 7 5.8 809 X 0.783
whence
sin /3 = 0.736 and /3 = 47.4°.
The velocity diagram may now be completed, and from it,
V2 = 1,290 and a = 30.5°.
This nozzle-exit velocity V2 is equivalent to 33.3 B.t.u., and

for this the value of 6„ is approximately 0.95 (p. 77),
hence qa = heat available for stage = 33.3 -^ 0.95 = 35.1
B.t.u.
To calculate the stage efficiency, the losses will first be
totaled. The nozzle loss = 35.1 - 33.3 = 1.8 B.t.u.
Since the heat equivalent of V^r = 890 is 15.8 B.t.u., and
of 73,= 809 is 13.1 B.t.u., the loss in the blading = 15.8 -
13.1 = 2.7 B.t.u. The leaving loss is 8 B.t.u. The leak-
age and disc friction loss is almost negligible at this stage:
0.3 per cent will be assumed, or 0.1 B.t.u. The total loss is
1.8 + 2.7 + 8 + 0.1 = 12.6 B.t.u., and the stage efficiency
6, = (35.1 - 12.6) -- 35.1 = 0.642.
The leaving loss for the second (from the last) stage will
be assumed to be 3 B.t.u. (This is taken less than the
leaving loss for the last stage because the blade angles will
be much smaller. If the assumption is in error, it can be
corrected later.) If the carry-over is 50 per cent, then the
heat drop for the last stage is 35.1 — 1.5 = 33.6 B.t.u.
This added to the heat content at the end point, 1 in. Hg
and 1.75 entropy, gives 941 + 33.6 = 974.6 B.t.u. for
the heat content of the steam leaving the second stage.
Constant entropy will be assumed for this calculation
because the work is then much simplified without affecting
the final result appreciably, when only a few rows are
calculated. At heat content 974.6 and entropy 1.75 the
pressure is 1.92 in. Hg and the volume 308 cu. ft.
236 STEAM TURBINES
For the second stage, the blade height and thickness
be assumed the same as for the last stage.
coefficient will
Hence the net area for axial flow = 543 sq. in. as before,
and the axial component V^f of the exit velocity F3 will be
..
Vsf = 4.26 X 308 X 144
= ^ .^
348
,.
ft. per second.
,
_-,o ^
543
The maximum flow for a given leaving velocity being
obtained when Vz coincides with F3/, the velocity diagram
ps^j.,^^^ in Fig. 219 is started
S^^^^».,^^ with V3 = 348 drawn
^^J?5»ir^"^"^'-^:i^ axially, and Vb = 503.
^:^ ^""~--.,,.,^^^^ The closing line of the
I^.!1^]_L^^ «^^g?^^^^ exit triangle is V^r =

'''^''
^
''-'''
612, and sin 7 = 348
^'''" ~^^'
-r 612 = 0.569, whence
7 = 34.7 deg. For a relative velocity of612 ft., kh = 0.95,
and V2r = 612 -- 0.% = 645 ft.
Following the procedure of the previous stage in regard
to nozzle-exit height and blade-entrance height,
6^ ^ 645 X s i n ^
5.8 612 X 0.569
whence
sin iS = 0.558, and /S = 33.9°.
The velocity diagram may now be completed, and from it
V2 = 1,100 ft. and a = 19.2°.
This nozzle exit velocity is equivalent to 24.1 B.t.u., for which

e„ = 0.965, hence qa = 24.1 in. -^ 0.965 = 25.0 B.t.u.
For this second stage the stage losses total 4.6 B.t.u.,
hence the efficiency e, = (25.0 - 4.6) -^ 25.0 = 0.816.
The leaving loss is 2.4 B.t.u. instead of 3 B.t.u. as assumed,
therefore the heat drop for the last stage should have been
35.1 - 1.2 = 33.9 B.t.u. instead of 33.6, and the heat
content would have been 974.9, very slightly raising the
pressure and decreasing the volume. While negligible in
this case, the discrepancy could be removed by recalculat-
ing with the new value of the leaving loss.
:
DESIGN 237
The nozzle exit velocity V2 = 1,100 being equivalent to

24.2 B.t.u. for which €„ = 0.96, the heat drop in the nozzle
= 24.2 --- 0.96 = 25.2 B.t.u. Assume the leaving loss for
the third stage to be 1.2 B.t.u. The stage heat is then 25.2
- 0.6 = 24.6 B.t.u. Hence 941 + 33.9 + 24.6 = 999.5
B.t.u. heat content. At this heat content and 1.75 entropy,
p = 1.49 lb. and v = 204 cu. ft.
The nozzle and blade-exit angles have been decreasing

from the last stage toward the earlier ones, and inspection
indicates that this third stage will be normal and similar
to those for the higher pressures in the turbine. Hence
the angles become ,^^
normal and the blade I
^^J^"----^^ ^
height is to be deter- Is '"^STjTr^^^
mined. Assume a = ^:<.-^ ^"""^^^
-^
1=-"
18° and Vt = 503: a \
/?= ^? ^<'
iA . cm
--^^"^-^
few trials indicate that Fig. 220.

with V2 = 1,070 ft. an
efficient relation can be found (Fig. 220) The stage leaving .
velocity is 284 ft. equivalent to 1.6 B.t.u. (instead of 1.2

assumed above), t = 30 deg. and V^r = 575.
The blade-exit height h^ (in inches) may be calculated
from the following relation
Volume of steam flow = X v, W
= /i3 X circumference X sin 7 X 0.9 X Vsr.
Hence
h - 4.26 X 204 X 12 X 12 _ .
~ X 0.5 X 0:9 "

^' ^'^^ '''•
5.75 X 32 X TT
This is less than 6 in., the allowable height, which justifies

our assumption that this and the remaining earlier stages
are normal that is, they do not need to have increased
;
angles and heat drops as is the case with the last two stages.
The design of the earlier stages could now be started
from the high-pressure end, using as the end-point pressure
that found above at the entrance to the stage second from
the last, 1.49 lb. The entropy 1.75 will probably be slightly
:
238 STEAM TURBINES

modified and the two last stages can be recalculated if
necessary.
The above problem illustrates the limiting case. If it
is necessary to have a steam flow greater than 4.26 lb. per
second, then additional area must be obtained by increasing
the blade height or the blade-ring diameter, or an increased
leaving loss must be accepted. Increasing the exit angles
further will give no help. If a less flowis required, then
the height of the blades may be reduced, or better, the

efficiencymay be improved by decreasing the exit angles
and the leaving loss. Also, the direction of V3 need no
Jonger be axial and some improvement may be obtained by
changing it.
88. Design Problem Velocity Stage -impulse Turbine.

:
Calculate the principle dimensions of the nozzles and

blading of a non-condensing turbine for the following
conditions
Kilowatts delivered at the shaft coupling 500

,
Maximum allowable blade speed 500

Exhaust pressure, pounds absolute 16
Throttling governing.
Turbine to be similar to Figs. 19 and 21 in general arrangement but with
only two pressure stages.
Part A. —Heat distribution, velocities, and stage pressures.

From the given conditions,
hi = 1,290.1 = heat content at 150 lb. and 180°;

Jh = 1,102.0 = heat content at 16 lb. and 1.676 entropy;
Qa = 188.1 = available heat.
For a non-condensing turbine with only two pressure

stages, a reheat factor R = 1.02 may be assumed. Hence
Qc = 188.1 X 1.02 = 191.8 = cumulative heat.

DESIGN 239
For each of the two stages, g„ = 191.8 -^ 2 = 95.9 B.t.u.

For this heat drop, e„ = 0.875 from Fig. 79, and the heat
effective in producing velocity = 0.875 X 95.9 = 83.8, cor-
responding to Vo = 2,048 ft. Using p = 0.22 from Fig.
146, Vb = 0.22 X 2,048 = 450 ft. per second. With 3,600
r.p.m. this gives
Circumference = 450 X 60 ^ 3,600 = 7.5 ft. = 90 in.
Diameter = 2.38 ft. = 28.6 in.
Draw the velocity diagram with Vb= 450, Vo = 2,048,

and a = 15 deg. The both moving and guide
exit angles of
blades will be reduced a moderate amount. The value of
72r = 1,612: for this velocity = 0.90 and Vsr = 0.90 X
h
1,612 = 1,451. For Fg = 1,035, A;^ = 0.88 and V, = 0.88
'if
—
240 STEAM TURBINES

second trial with g„ = 97 and g, = 97 X (1 - 0.66) =
32.96 B.t.u. gives the correct pressure, as follows:
Table XII. Stage Heats for Two-stage Turbine
Stage
DESIGN 241
The efficiency calculated with the revised velocities, by

Eq. (61a), is 0.805, the same as before.
The total work done per pound of steam = Ei = 2 X
(97 - 32.96) = 2 X 64.04 = 128.1 B.t.u. Therefore,
ei = internal efficiency = 128.1 ^ 188.1 = 0.682

R = reheat factor = (2 X 97) ^ 188.1 = 1.031
or
= 0.682 ^ 66 = 1.033.
There will be no difficulty in getting sufficient steam

through the last rows of blading of this small non-condens-
ing turbine, hence the leaving loss may be calculated from
the velocity F5 = 226, from Fig. 222. This is equivalent to
1 B.t.u., or 1 -^ 188.1 = 0.5 per cent. The losses are taken
as follows:
Bearing friction 2.5

Shaft-gland leakage 2.0
Radiation 2.0
Pump and governor drive 2.0
Leaving loss 0.5
Total 9.0 per cent
Deducting this from the internal efficiency,
66 = 0.682 - 0.090 = 0.592.
The steam per kilowatt-hour for the ideal cycle = 3,413 -f-
188.1 = 18.15 lb. The steam per kilowatt-hour delivered

at the coupling = 18.15 -- 0.592 =30.65.
Part B. —Dimensions of nozzles and blades.
Stage 1.— The total steam used is 30.65 X 500 =
15,325 lb. per hour or 4.28 lb. per second.
The stage 1 nozzles expand from 150 lb. to 54 lb., the
critical pressure being po = 150 X 0.577 = 86.5 lb. At
<p = 1.676, ho = 1,235 B.t.u. and Vo = 5.90 cu. ft. The
heat drop to the throat = go = 1,290.1 - 1,235 = 55.1
B.t.u., the equivalent velocity Vo = 1,660.
, 4.28 X 5.90 X 144

= ^ ,„ , , ,, ,
Ao =
. ,
^r-r^r, 2.19 sq. m., total throat area.

IjObU
— :
242 STEAM TURBINES
At exit, hi = 1,193.1 B.t.u., V2 = 8.37 cu. ft., and the total

heat drop = 1,290.1 - 1,193.1 = 97.0, the velocity being
V2 = 2,061 as already found. Hence
A =
,
2
4.28 X 8.37
--
X 144
= ^ -^, ^ ^ ..,
2.501 sq. m., total exit area.
,
^
2,061
The blade heights should not be than l}i to 2 per less

cent of the blade-ring diameter, 28.6 or between 043 in.,
and 0.82 in. Round section reamed nozzles will be used.

Trying 10 nozzles, the exit area of each = 2.501 -^ 10 =
0.2501 sq. in., corresponding to 0.564 in. diameter of nozzle
exit. Adding 0.1 in. for excess of blade entrance height
over nozzle exit height gives 0.664 in. which is within the
limits above.
The throat area = 2.19 -^ 10 = 0.219 sq. in., corre-
sponding to .528 in. dia. at throat.
With a coefficient of 0.90 to allow for the wall thickness
between the nozzle at exit,
j^.. , diameter 0.564 ^ .^^ .
and 10 nozzles occupy 24.2 in. of arc, or 24.2 -^ 90 = 0.27

of the complete circumference.
Since the blade-entrance angles will be as found from
the diagram, the necessary height at entrance to the first
moving row wdll equal the nozzle exit height, (p. 127), if
the same thickness coefficient is taken for the blades. The
various heights will be as follows (Eq. (71))
Table XIII. Blade Heights

Incli
Height first moving row at entrance = . 664

542
h^ = height first moving row at exit = 0.564 X jj^ = .695
height guide blade at entrance = . 79

542
hi = height guide blade at exit = 0.564 X 0.950
^22
height second moving row at entrance = 1 . 05
542
hb = height second moving row at exit = 0.564 X ôa ^ ^ ^^'^
DESIGN 243
The height ratio = 1.353 ^ 0.564 = 2.4. This rule is

probably based on square section nozzles, however, hence
the ratio should be 1.353 -^ \/.25 = 2.72, which is between
the limits given on page 132.
The nozzle and blade heights in Table XIII were cal-
culated without allowance for the increase of volume due to
the reheat occurring during 1300
the flow through the pas-

sages, and since there is con-
siderable reheat due to
the high steam velocities,
it is worth while to digress
for a moment to deter-

mine the magnitude of the
effect.
The nozzle efficiency
being €„= 0.875, the nozzle
reheat = (1 - 0.875) X 97
= 12.1B.t.u. Adding this
to the heat content at the
nozzle exit from Table XII,
1,193.1 + 12.1 = 1,205.2
B.t.u. = true heat content
at the nozzle exit, and at 54
lb. the volume is 8.69 instead
of 8.37. Therefore the exit
area should be increased to
0.25 X 8.69 -^ 8.37 = 0.262,
and the diameter = 0.578
in., an increase of 0.014 in.
This is insufficient to affect the blade-entrance height

adopted above.
The blading efficiency is 0.805, therefore the loss in the
blading is 0.195. The steam entering the blades at Vo =
2,061 has kinetic energy expressed in B.t.u. = 97 X 0.875
= 84.9, therefore, the loss in the blading = 84.9 X 0.195
= 16.6 B.t.u. Of this, 1 B.t.u. is due to exit velocity F5,
244 STEAM TURBINES
hence the true reheat in the blade passages is 16.6 — 1.0
= 15.6 B.t.u. This added to the heat content at the
nozzle exit as previously found (= 1,205.2) gives 1,220.8
B.t.u. as the heat content of the steam leaving the last row
of blading. At this heat and 54 lb. pressure the volume
has increased to 9.06 cu. ft. This would increase the exit
height of the last row in the ratio 9.06 -^ 8.37 or 1.083, and
h, = 1.353 X 1.083 = 1.466 in. New values of hz and h^
can be estimated similarly or they can be interpolated.
The above apparent deficiency in flow area through the
blades is by no means negligible. On page 132 there was
mentioned, however, a possible gain from such a reduction
of area for the flow through three rows. Neglecting the
reheat in the above problem is equivalent to a reduction
of 1 - (1 -- 1.08) = 7.4 per cent.
Figure 223 shows on the Mollier chart the first stage

heat drop and the various reheats determined above.
—
Stage 2. The stage 2 nozzles expand the steam from
54 lb. to 16 lb., the critical pressure being po = 54 X 0.577
= 31.15 lb. At ^ = 1.716, ho = 1,179.4 B.t.u.; and Vo =
13.92. The heat drop to the throat = Qo = 1,226.1 -
1,179.4 = 46.7 B.t.u., the equivalent velocity Vo = 1,528.
Stage leakage can be taken into account here. Assuming
a 3-in. shaft, five rings of labyrinth packing and 0.020-in.
clearance, the line for five rings on Fig. 184 may be extended,
giving a flow of 250 lb. per hour. Correcting this,
Leakage = 250 X ^^ ^ m ^ ^^^ '^* P^"" ^'''^^'
This is 270 ^ 15,325 = 1.76 per cent. Deducting this from

the steam flow, 15,325 — 270 = 15,055 lb. per hour through
the stage 2 nozzles, equivalent to 4.18 lb. per second.
Hence
4.18 X 13.92 X 144
= 5.48 sq. m.
1,528
4.18 X 24.2 X 144
= 7.066 sq. in.
2,061
DESIGN 245
It is desirable if possible to use a duplicate of the first-
stage wheel in the second stage. Therefore
Number of nozzles = 7.066 - 0.2501 - 28.25.

Use 28 nozzles.
Exit area = 7.066 -^ 28 = 0.255 sq. in. corresponding to

0.570 diameter instead of 0.564 as before, a difference
in.
of 0.006 in. This would require an increase of (0.570 —

0.564) -^ 0.564 or only 1 per cent in height, and the same
wheel design will suffice.
The throat = 5.48 -^ 28 = 0.1957 sq. in. area or 0.499
in. diameter.
The nozzle pitch = 0.570 -- (0.2588 X 0.9) = 2.445 in.
and 28 nozzles occupy 28 X 2.445 = 68.5 in. of arc or 68.5

-^ 90 = 0.76 of the circumference.
To calculate the disc friction losses, the following data
are required: d = 28.6 in., Vi = 450, average blade height
= 1 in., c = 1.25, m =
1, specific volume = 9.2
1.1, n =
and 25.6 and 2 respectively, and per cent of
for stages 1
active periphery = 27 and 76 respectively. Substituting

in Eq. (84),
^ , , ^ /450V V
Viooy
= 14.75 for the
28.(
9.2 So +
first stage.
^' - '-''^ X W
Similarly hp. = 2.64 for the second stage.
Total hp. = 17^39 hp. = 13 kw.
Summary. —The steam supplied = 15,325 lb. per hour.
Of this
2.0 per cent is lost by radiation.

2.0 per cent is lost by gland leakage.
0.9 per cent is lost by stage leakage (average of 1.76 for
two stages)
4.9 per cent total.
15,325 X (100 - 4.9) = 15,325 X 0.951 = 14,570 lb.
effective.
246 STEAM TURBINES
Since the theoretical water rate is 18.15 lb. per kilowatt-

hour this should produce 14,570 ^ 18.15 = 803 kw.
The combined nozzle and blade efficiency is 0.70, so that
803 X 0.70 = 562.1 kw. is the work delivered by the
blading. Deducting the work of disc friction,
562 - 13 = 549 kw.
Allowing 2.5 per cent for bearing friction and 2 per cent
for governor and pump drives, or a total of 4.5 per cent,
549 X 0.955 = 524 kw. delivered to the coupling.
This is nearly 5 per cent in excess, and if it must be reduced,

recalculation is necessary.
89. Reaction Turbine Design, General Considerations.
The design of a reaction turbine is somewhat more

involved than that of an impulse turbine, because of the
following features:
1. There is necessarily full peripheral admission in all of
the turbine stages.

2. The stages are usually much more numerous.
3. The diameter at the high-pressure end is made as
small as possible, and that at the low-pressure end
as large as possible.
4. The progression in blade-ring diameter and blade
height from the high-pressure to the low-pressure
end is not uniform, but is subject to variations of
three sorts:
A. The diameter of the rotor on which the blades
are placed is varied in three or more steps
called drums, the relative diameters being
more or less arbitrarily selected.
B. The blades on each drum increase in height by
groups termed expansions, there being two
to seven expansions per drum, and a number
of blade rows in each expansion.
DESIGN 247
C. The mean blade-ring diameters of a series of

expansions may be constant, as in Fig. 27,
giving a constant blade speed for a whole
drum, or instead, the drum may be of con-
stant diameter, with the increase in height at
the expansions taking place outwardly, as in
Fig. 29. In this case each expansion has a
different mean blade speed.
In addition to the above, provision is often made for the

increase in steam volume within each expansion, by
gaging, the first stages being slightly closed, and the later
ones slightly opened. If the changes in blade height are
sufficiently frequent,however, this is not done.
Reaction turbines are variously arranged and propor-
tioned. The straight, high-pressure condensing type with
three drums, as illustrated in Fig. 27, might be termed the
fundamental type. It is common to have the diameters
of the consecutive steps vary in proportion to V'2, or
high: intermediate: low = 1:1.414:2. The blade speeds
vary in like proportion, also the appropriate steam speeds,
and hence the heat drops per stage will vary as the squares,
or as 1:2:4. With
this proportion, the work is usually
divided among
the three drums in the proportion 1:1:2.
Variations of this are not uncommon; for example, a
shorter turbine maybe built by making the low-pressure
drum do more work, and a work division of 1 2 4 has been
: :
used with, of course, appropriate variation in the diameter

ratios. Also, more than three steps may be used. If the
work is divided between two cylinders in a tandem or
cross-compound unit, other arrangements may be used;
in Fig. 42 the high-pressure cylinder has two drums and
the low-pressure also two. Furthermore, in all except the
largest reaction turbines, it is customary to replace the
high-pressure blading by a two-row impulse wheel, and
this may be said to decrease by one the number of drums of
reaction blading. In some cases, where maximum low-
248 STEAM TURBINES
pressure flow area is desired, the diameter of the low-

pressure drum has been made more than twice that of the
high-pressure drum.
To gain area at the low-pressure end for the large volume
of steam, the low-pressure part often is made double flow,
as in Figs. 38, 40 and 41, a separate double-flow turbine

may be used as in Figs. 42 and 44, or multiple-exhaust
blading used as in Figs. 35, 36, and 37.
On the other hand, with the high-pressure reaction
blading replaced by a two-row impulse wheel, only one
reaction drum may be used, as in Figs. 30, 31, 33, and 39.
These cuts also illustrate conical bored cylinders (see also
Fig. 35) and angular tipped blading (see Fig. 161) instead
of the square-ended blading illustrated in most cases.
The rectangular step from one diameter to the next larger
may be replaced by a cone, as in Figs. 35, 37 and 39.
It will be noted that in most cases the last few rows of
blades are of the same height; as the volume of steam
increases rapidly here, increased exit angles must be used,
as in the case of the impulse turbine. These are often
called wing blades.
account of the large number of stages in a reaction
On
turbine, it would be very laborious to carry the design
through step by step as was done in the impulse turbine,
hence the blades are usually dealt with in groups, each
expansion often being treated as a unit.
The adoption of regenerative feed heating, by bleeding
steam from the turbine at one or more points to be used in
heating the boiler feed, has been of notable assistance to
turbine designers, in removing a considerable fraction of
the steam before it reached the low-pressure stages, and
therefore allowing the remainder a chance to expand fully
in the area obtainable. Practically all large units are now
operated in this manner, and many moderate size and small
ones also. While provision for bleeding has some effect on
the design of a steam turbine, tending to increase the flow
area provided in the earlier stages or decrease it in the
DESIGN 249
later ones, or both, the problem of regenerative feed heating

is really a part of power-plant design and is outside the
field of this text.
90. Design Problem Reaction Turbine.

:
Calculate the principal dimensions of a reaction turbine

to suit the following conditions :
Kilowatt delivered to shaft coupling 5 000

,

,
Blade speed, maximum .'
570
Exhaust pressure, inches absolute 2
Turbine to have three drums, with diameters approximately in ratio
1: \/2:2, and to have work distributed approximatelj' in ratio 1:1:2.
Blading on each drum to have constant mean diameter.
The data has been taken much like that of the problem in
Sec. 86, partly for the sake of comparison, and partly so
that some of the same data may be used.
The speed is rather high for reaction blades, but wdth
modern construction and materials it is allowable. The
same is true of the superheat; the older types of reaction
turbines were not supplied with highly superheated steam
on account of its tendency to loosen blade fastenings,
though a high-pressure impulse wheel preceding the
reaction blading could receive superheated steam and
lower it to a temperature satisfactory for reaction blading.
Modern materials and types of construction obviate this
difficulty,however, and if necessary, high-temperature steam
may be supplied directly to reaction blading, as is done in
the case of Fig. 42. It will be used in the present problem.
From pages 225 and 226 may be taken the following:
Low-pressure mean blade-ring diameter = 4.53 ft. =
54.4 in.
Circumference = 171 in.
Maximum allowable height low-pressure blading 10.9 in.
From the diameter ratios given in the data we derive,

High-pressure mean diameter = 54.4 in. -^ 2 = 17.2 in.,
250 STEAM TURBINES
Circumference = 85.5 in. High-pressure blade speed
= 570 -^ 2 = 285.
In accordance with data on page 147, p will be taken at
0.80, giving a steam speed V2 = 285 ^ 0.80 = 357, which
corresponds to 2.55 B.t.u. Assuming a nozzle efficiency of
0.90 (page 145), the heat drop in the first row is 2.55 -^
0.90 = 2.83 B.t.u.
Draw a velocity diagram with Vb = 285, V2 = 357,
a = 20 (see page 139). From this diagram, assuming 50
per cent reaction and therefore equal triangles, V3 is found
to be 133 ft. per second, corresponding to 0.33 B.t.u.
If it is assumed that 50 per cent of this is effective as carry-
over, then only 2.83 — 0.17 = 2.66 B.t.u, drop is necessary
in the fixed blades. With 50 per cent reaction, the heat

drop in the moving blades will also be 2.66 B.t.u., or the
total drop per stage (one fixed and one moving row^) will
be 5.32 B.t.u. This is for the high-pressure drum.
The efficiency of the blading is found from Fig. 155 to be
0.895. The combined efficiency of nozzles and blading is
0.90 X 0.895 = 0.805. If the average tip clearance leak-

age be assumed at 3 per cent, then the stage efficiency =
es = 0.805 X 0.97 = 0.78.
A table should now be made up
like Table VIII, with the
above stage but with arbitrarily chosen heat
efficiency,
drops per stage of 20 B.t.u. Pressures and volumes after
reheat should be noted. This data is shown in Table XV,
except that for convenience e^ has been taken at 0.80.
(The pressures and volumes are little affected by a small
change of e^.) From the heat content at the point nearest
2 in. abs. pressure, it will be found that the reheat factor is
DESIGN 251
1.04. The internal efficiency is therefore 0.78 X 1.04 =

0.811, exchiding the steam which leaks past the dummy
pistons.
Other losses are assumed as follows: dummy leakage, 3
per cent; radiation and bearing friction, each per cent; 1
water-gland loss and governor and pump work each 3^^ per
cent, or a total of 6 per cent. This leaves 0.811 — 0.06 =
0.751 for the brake efficiency. Use 0.75.
The steam consumption may now be calculated.
hi = 1,290.1 = heat content at 150 lb. and 180°.

/i2 = 934.8 = heat content at 2 in. Hg and 1.676 entropy.
qa = 355.3 = available heat.
Theoretical steam per kilowatt-hour = 3,412 -^ 355.3 =

9.61 lb.
Actual consumption = 9.61 -i- 0.75 = 12.8 lb.

12.8 5,000 X
-1.1 . .
— = ,^ô
• • • 1 ,1-
ihis IS equivalent to r7r\7^^n 17.78 lb. per second
= 64,000 lb. per hour
It is convenient to find the number of stages there would
be if all were on the same mean diameter as the high-pres-
sure stages. There would be 369.4 -=-
5.32 = 69.5 stages.
Since the division of work is in the proportion 1:1:2,
one-half of the above stages or 34.75 stages should be the
equivalent of the actual low-pressure stages, on a diameter
twice as great, and therefore wdth a heat drop four times
as great. Thus 34.75 ^ 4 or 8.69 stages would be required.
Nine stages will be used, with a heat drop of 4 X 5.32 =
21.28 B.t.u. per stage or a total of 9 X 21.28 = 191.52,
which would be the equivalent of 36 of the previously cal-
culated stages on the high-pressure diameter, leaving 33.5
stages to be divided between the high-pressure and inter-
mediate-pressure. Taking 16 stages for the high-pressure
with qa = 5.32 B.t.u. each or a total of 16 X 5.32 =
85.12, leaves 17.5 stages for the intermediate-pressure drum.
Since the intermediate-pressure drum is \/2 X high-pres-
sure diameter, the heat drop per stage of the intermediate-
—
252 STEAM TURBINES

pressure will be twice that of the high-pressure, and the
number of stages, one-half, or 8.75 stages, each with 2 X
5.32 B.t.u. = 10.64 B.t.u. drop. Evidently nine stages
would be used, each with (17.5 X 5.32) -^ 9 = 10.34
B.t.u. drop or a total of 9 X 10.34 = 93.06. The blade
speed might be adjusted to suit the new steam speed, and
would be
jlO.i
intermediate blade speed = 285 X V2 X \/i7f7 398.1
.64
and the corresponding diameter is
^^^^-^ -
~ 3 183
^-^^^ ft
^^• }8.20 in.
2,400 X TT
This may be compared with 27 X V2 = 38.31 in. which

would be the unadjusted diameter.
The 16 stages of the high-pressure drum will be divided
into 4 expansions of 4 stages each; the 9 stages of the
intermediate-pressure drum will be divided into 1 expansion
of 3 stages and 3 expansions of 2 stages each; and the 9 stages
of the low-pressure drum will each be treated as a separate
expansion. The last is often necessary because the increase
of volume is so rapid that continuous increase of blade
heights is The heat drops per expansion and
advisable.
per drum may now be calculated.
The data thus far accumulated will be assembled in
Table XIV for convenient reference.
Table XIV. Heat Distribution

—
DESIGN 253
The pressures of the steam at the changes in drum diam-

eter, and the volumes at the ends and centers of each
expansion must now be obtained in order to calculate blade
heights. The starting point for this is Table XV.
Table XV. Data for Cumulative Heat Diagram
Heat drop
254 STEAM TURBINES
400 400
0.5 1 5 45 iO.O 10 50 4050 100 200 300

Pressures Log. Scale
Fig. 225.
:
DESIGN 255
A and C2, etc., measured in B.t.u., from an expansion curve

like one of those in Fig. 193, by e^, the stage efficiency.
With the data of Table XV at hand, however, it is
simplest to take 1 X 20, 2 X 20, 3 X 20, etc. (as entered
in column 6, Table XV), as the successive values of Qc
and lay them off as ordinates with the corresponding
pressures as abscissae. A logarithmic pressure scale is
used, however, in order to open out satisfactorily the lower

part of the scale. The curve thus obtained in Fig. 225
is called the cumulative heat curve. The volumes from
the table are also plotted, and for the sake of accuracy the
smaller volumes are replotted to a tenfold scale.
The ordinate cumulative heat curve at the final
of the
pressure, 2 in. = 0.982 369.4 B.t.u. This is divided
lb., is
into three main divisions corresponding to the total heat

drops of the high-pressure, intermediate-pressure, and low-
pressure drums, and each is further subdivided into por-
tions corresponding to the heat drops of each expansion.
(This subdivision could be extended to include each stage
if desired.) Horizontal lines are drawn to intersect the
cumulative heat curve, and verticals to the pressure scale
determine the pressures. At the intersection of these
verticals with the volume curve may be read off the volumes.
A line representing the center of each expansion is also
drawn.
Table XVI is the result of calculations for blade heights
and gaging, made from data taken principally from Fig.
225, the cumulative heat diagram. Data in the various
columns were obtained as follows
—
Column 3. Values may be read off the curve, or easily
calculated and are included merely for completeness.
—
Columns 4 and 5. Values read from the diagram.
—
Column 6. The steam velocities, and blade-ring diam-
eters and circumferences have been previously deter-
mined. The steam flow is determined as follows:
While the total amount of steam to be supplied to the
turbine has been found to be 17.78 lb. per second or 64,000
—
256 STEAM TURBINES

Table XVI. Reaction Blade Calculations
s
3
P
:
DESIGN 257
lb. per hour, it is necessary to deduct the steam leaking

past the dummy pistons before the amount passing through
the blading may be determined.
The purpose of dummy pistons was briefly stated on
page 157. To determine the exact diameter necessary to
accomplish this requires a knowledge of the blade heights
which is not available at this point in the calculations.
It will therefore be assumed that the dummy diameter is
equal to the mean blade diameter of the corresponding drum
(Fig. 27). The leakages in pounds per hour may be read
for 0.010 in. clearance and 100 lb. pressure from Fig. 184;
see also page 175. The weights read from the chart will be
modified as follows
High-pressure dummy, diameter = 27.2 in., pressure
= 1501b.
Assume 20 rings and 0.012 in. clearance.
Leakage = 1,210 X ~ X ^^ = 2,180 lb. per hour.
= 2,180 ^ 64,000 = 3.41 per cent.
Steam passing high-pressure blading = 64,000 — 2,180

= 61,820 lb. per hour = 17.17 lb. per second.
I ntermediate-pressure dummy, diameter = 38.2 in., pres-
sure = 62 1b. (Table XVI).
Assume 15 rings and 0.012 in. clearance.
Leakage = 2,100 X tttt^ X Q^vrn = 1,561 lb. per hour.
= 1,561 -- 64,000 = 2.44 per cent.
Steam passing intermediate-pressure blading = 64,000

— 1,561 = 62,439 lb. per hour = 17.35 lb. per second.
Low-pressure dummy, diameter = 54.4 in., pressure = 19.6
lb.
. Assume 10 rings and 0.012 in. clearance.
Leakage = 3,600 X ^qq X ^^ = 846 lb. per hour.
= 846 -^ 64,000 = 1.32 per cent.

258 STEAM TURBINES
Steam passing low-pressure blading = 64,000 — 846 =

63,154 lb. per hour = 17.55 lb. per second.
The average leakage, allowing for the fact that the low-
pressure drum does about twice as much work as either of
the others, is
(3.41 + 2.44 + 2 X 1.32) 4- 4 = 2.12 per cent.
—
Column 7. Assuming 25 per cent gaging at the middle of
each group on the high-pressure and intermediate-pressure
drums,
W
X volume X 144
^ ~ velocity X circumference X 0.25
This is the blade height for the whole expansion. On the

low-pressure drum the volume at exit from each stage is
taken instead of that at the center, since each expansion

contains only one stage, or one fixed and one moving row.
In this design, the mean diameter has been kept constant
on each drum. If the drum diameter, at the root of the
rotor blading, is kept constant, then the mean blade
diameters increase from expansion to expansion, requiring
a greater heat drop to accompany the increased blade
speed. The distribution for this case may be effected by
means of Fig. 225 also, but several trials may be necessary
to get a satisfactory result. Equation (79) is useful in
that case.
Column 8. —The leakage efficiency may be found from
Fig. 185 for each of the various lengths of blades and rotor
diameters. The average leakage efficiency is seen to be
about 0.965.
Column 9. —The ratio expansion the ratio
of is of the
volumes at the beginning and end an expansion.of
Column 11. —The blades preceding the center of an
expansion must have a smaller gaging, and those following
it a greater gaging than 25 per cent. For the first high-
pressure expansion, the ratio of expansion is 1.185, which is
the increase after going through four stages. The expan-
sion per stage is therefore \/1.185 = 1.043. The gaging
DESIGN 259
of stage 2 might be smaller than 0.25 in the ratio 1 : 1.0215,

and that of stage 3 larger in the inverse ratio ; for simplicity,
the second stage of each expansion (where there are four
made 25 per cent, and the others calculated
stages) will be
from it. In expansion 1 of the intermediate-pressure
drum, the gaging varies in the ratio v^ 1.325 and the middle
stage is made 25 per cent gaging. Where there are two
stages per expansion, the first is made 25 per cent 'gaging
and the second one larger in the ratio \/r.
The italic numbers in column 7, stages 7 to 9 of the low-
pressure drum, are the calculated heights. Since the
heights are limited to 10.9 in., the gaging is increased,
as shown in column 11. These values are calculated the
same way as were the last two rows of the impulse turbine.
The problem may be finished, output calculated, etc.,
following a method similar to that used in the preceding
problems.^
^ Those wishing undertake a more accurate calculation of the blading

to
an excellent method explained in Sterling's
of a reaction turbine will find
"Marine Engineers' Handbook," in the portion on Steam Turbines written
by A. G. Christie, McGraw-Hill Book Company, Inc.
PROBLEMS
Constants, coefficients, etc., necessary for the solution
of the. followingproblems and not included in the state-
ments, are to be selected from the text, proper considera-
tion being given to the particular conditions in each case.
CHAPTER II
1. A steam nozzle is to expand 2 lb. of steam per second from an initial
pressure of 200 lb. absolute per square inch to a final pressure of 80 lb. The
steam is initially superheated 200°F.
Calculate the cross-sectional areas at the throat and the exit, and the exit
velocity, for
(a) Frictionless flow.
(b) Flow with friction.
Also calculate, for (6),
(c) The final heat and entropy.
(d) The nozzle efficiency.
(e) The nozzle efficiency if the nozzle is cut short so that there is 15 per
cent underexpansion.
2. Steam having a pressure of 80 lb. absolute and a quality of 97 per cent
isexpanded through a nozzle of round cross-section 2.50 sq. in. in area at its
smallest section, into a space where the pressure is 55 lb. absolute. Making
suitable allowance for friction, calculate:
(o) Steam flow in pounds per second.
(6) Exit velocity.
Entropy, quality, heat content, and volume of steam at exit.
(c)
3.the steam entering the nozzle in Problem 2 has a velocity of 400 ft. per
If
second due to exit from the blading of a previous stage, calculate the same
quantities called for in Problem 2.
4. At the diaphragm of an
exit of the nozzles in the last low-pressure
impulse turbine, the total cross-sectional area available for flow is 1,000 sq.
in. The required steam flow is 16 lb. per second, the specific volume of the
steam at the nozzle exit is 550 cu. ft. per pound and the absolute pressure is
1 in. Hg. What is the necessary heat drop and what initial pressure is
required for the nozzles of the stage?
5. If, in Problem 4, the steam enters at a velocity of 300 ft. per second
instead of having zero entrance velocity, what will be the necessary heat drop
and the initial pressure?
6. Compare the throat velocity calculated in Problem 1 with the velocity
of sound under the conditions existing at the throat.
260
PROBLEMS 261
7. If in 1, the cross-section at the throat is square and the cross-

Problem
section at exit rectangular (Fig. 131), with the height the same as the
is
height of the section at the throat, calculate the throat and exit dimensions.
Also, if the included angle of divergence is 12 deg., find the length of the
divergent portion of the nozzle. Finally, if a series of these nozzles is placed
in a row along an arc, the thickness of the partition between them being 0.10
in., what will be the value of m, the thickness coefficient, and what will be
the pitch (Fig. 86) of the nozzles along the arc?

8. The nozzles in a diaphragm are made up like Figs. 86 and 87 by casting
curved partitions of sheet metal into the metal of the diaphragm, the exit
angle being 20 deg., the thickness of the partitions at the edge being 0.08 in.
and their pitch 2.76 in. There is a complete ring of nozzles and the mean
diameter is 54 in. If the steam flow is 16 lb. per second, the specific volume
of the steam at exit is 148 cu. ft. and the exit steam velocity is 1,390 ft. per
second, find the required radial height of the nozzles at exit.
CHAPTER III
Given the velocity of impulse blading = 520 ft. per second, the nozzle
1.
angle = 17 deg., kb = 0.91, and equality between blade entrance and exit
angles, draw to a scale of 20 ft. per inch three forms of velocity diagrams and
find the velocity of the steam at exit from the nozzle, the force in pounds
exerted on the blades and the work done in B.t.u. per pound of steam flow,
and the efficiency of the blading (the last by three methods),
(a) Whenthe speed ratio p = 0.43.
(6) Whenthe speed ratio p = 0.52.
2. Repeat Problem 1 with the same data except that the blade exit angle
is made
8 deg. less than the blade entrance angle.
3. A is to be designed for the following con-
single-stage impulse turbine
ditions: initial steam pressure = 140 lb. absolute, superheat = 50°F.,
exhaust pressure = 16 lb. absolute, nozzle angle oc = 20 deg., blade exit and
entrance angles equal, and blade speed = 1,000 ft. per second. Find fc„
from Fig. 79, take kb = 0.88, draw the velocity diagram, calculate nozzle
efficiency, blading efficiency, and combined nozzle and blading efficiency.
Assume blade width = î in.; find pitch and draw, to twice their size, the
sections of three consecutive blades.
4. In Problem 3 what will be the horsepower delivered by the blading if
the steam flow is 3 lb. per second? Also, what will be the total area required
at throat and exit of the nozzles?
5. In Problem 4, calculate the number of nozzles, the nozzle dimensions,
and the blade dimensions under the following conditions: The nozzle
section at the throat is approximately 0.60 in. square, the nozzle height at
exit is the same as at the throat, the blade height at entrance is 0.12 in.
greater than the nozzle height at exit, the thickness of the nozzle partitions
at the exit edge is 0.075 in., and of the blade edges is 0.020 in. Also find
the pitch of the nozzles and, using the blade width and pitch found in
Problem 3, find the height of the blades at exit.
262 STEAM TURBINES
6. In Problem 5, suppose that the blade entrance and exit angles are
arbitrarily made 35 and 24 deg., respectively, the same nozzle proportions
and Fj being used. Draw the velocity diagram and calculate
exit velocity
the blade entrance and exit heights. What heights should be used?
7. With the same initial and exhaust conditions for the steam as in
Problem 3, the same nozzle angle and S3^mmetrical blades, draw the velocity
diagram for a two-row velocity-stage turbine, taking p = 0.20, and finding
ki from Fig. 140. Calculate the blading efficiency, the combined nozzle and
blading efficiency, and the horsepower, with a steam flow of 3 lb. per second.
Draw three blade sections in each row, selecting appropriate blade widths
from Sec. 47. Assuming a nozzle height of 1 in. at exit, calculate the exit
heights ofmoving and guide blades.
8. Repeat Problem 7 with the same data except that the exit angles for
blades and guides are decreased a suitable amount.
9. The maximum allowable blade speed two-row wheel
for a given type of
construction is 600 ft. per second. Using =16°, reducing the blade exit
oc
angles and selecting suitable values of p and h from the text, draw a velocity
diagram and calculate the blading efficiency. Also calculate the heat drop
for the nozzle supplying this wheel, assuming Vi = 0.
10. Calculate the steam flow in pounds per second through a ring of
fixed reaction blading having a mean diameter of 30 in., blade height =
0.87 in., and blades gaged 25 per cent. The steam is at 40 lb. absolute
pressure and is superheated 10°F. on the inlet side, and is at 31 lb. absolute
pressure on the discharge side. Make the calculation,
(a) Neglecting flow through the clearance area.
(6) Finding clearance and allowing for leakage through clearance area
(Sees. 60 and 65).
11. In a certain stage of a reaction turbine the blade speed is 400 ft.
per second, the steam speed (F2 = V->r) is 550 ft. per second and the exit
angle of the blades is 20°. Draw the diagram, find /3, calculate the heat drop
per stage, the work done per pound of steam and the blade efficiency,
(a) With €co = 0.
(b) With =
50 per cent.
eco
12. Sketch a reaction blade '2 in. wide having entrance and exit angles
of 85 and 20 deg., respectively, and with convergent exit. Make two tem-
plates from this blade, on tracing cloth or in celluloid, and by manipulation
of these templates show the effect of change in pitch and angle on gaging
and degree of convergence, as follows:
(a) With constant angle, show change. in gaging and convergence or diver-
gence at exit, as the pitch is changed.
(b) With constant pitch, show change in gaging and convergence or diver-
gence, at exit as angle is changed.
CHAPTER IV
1. Calculate the leakage through dummy
packing in pounds per square
inch of leakage area per second, for the case where steam is at an absolute
pressure of 190 lb. per square inch and is superheated 150°F., the pressure
at the other end of the packing being 50 lb. absolute. Calculate for 5, 10,
—
PROBLEMS 263
15, and 20 duininy rings. What is your conclusion as to the effect of

increasing the number of dummy rings?
2. Smooth discs are rotated in air at a peripheral speed of 600 ft. per
second. Compare the frictional losses for discs of different diameters
rotating at 1,200, 1,800, 3,600, and 4,800 r.p.m.
3. Explain the theory of operation of a water-seal gland at the exhaust
shaft packing of a turbine.
CHAPTER VII
Design Problems 1 to 9, for which data are given in the table and the
succeeding descriptions, cover a range of sizes and types sufficient to illus-
trate the methods explained in the text for the calculation of steam rate,
nozzle and blading dimensions, and arrangements, etc. Problems 10 and 11
illustrate the preliminary determination of the final stage dimensions and
limitations; such predetermination should logically precede the solution of
any of the design problems, but the average reader will find them easier
after he has had the experience of going through one of the earlier problems.
Table XVII. Data for Problems 1 to 9

264 STEAM TURBINES
7. Combination impulse and reaction turbine, two-row impulse wheel
followed by two drums of reaction blading.
8. Same as Problem 7 except that the impulse wheel is followed by a
conical drum of reaction blading (Fig. 30).
9. Impulse turbine, single-wheel re-entrj' t3'pe, throttling control, sym-
metrical blading (Figs. 23 and 24).
10. Find the maximum possible steam flow in pounds per second for a
leaving loss equivalent to 12 B.t.u. in a turbine rotating at 1,800 r.p.m. with
a maximum blade speed at mean blade height of 600 ft. per second. The
maximum allowable blade height is 20 per cent of the mean blade-ring
diameter, the final pressure is 1 in. Hg, and the final entropy is 1.78.
11. Make a provisional calculation of the low-pressure stages of the
turbine in Problem 10, starting at the last and proceeding backward until
normal conditions are reached, these being in this case, a nozzle angle of 18
deg. and with blade exit angles decreased from the values at blade entrance.
Blade entrance and exit heights are to be equal.
12. If the allowable blade speed in Problem 10 is increased 10 per cent
by improved material and construction, thus allowing greater diameter
and higher blades, by what per cent will the maximum possible steam flow
be increased, other conditions remaining the same?
13. If the maximum allowable steam flow found in Problem 10 for a leav-
ing loss of 12 B.t.u. be increased 15 per cent (by by-passing steam at
admission), the entropy increasing to 1.80 but other conditions remaining
the same, what will be the leaving loss?
14. The following is the data of the last five rows of blading on a 3,000-kw.
reaction turbine:^
Row number
(from last)
PROBLEMS 265
16. A small single-disc turbine of the re-entry type has two separate
nozzles, one of which operating alone will develop 35 brake horsepower when
supplied with dry saturated steam at 150 lb. gage pressure, the back pres-
sure being atmospheric. The two nozzles operating together under the same
steam conditions will develop 50 brake horsepower. The turbine rotor is
17 in. mean
diameter, has blades 1 in. wide and %
in. high, and runs at
7,200 r.p.m.; the nozzle angle is 173^ deg., the blade inlet and outlet angles
are both 23 deg. and the reversing chamber exit angle is 15 deg.
(a) How much power will the small nozzle develop when operating alone?
(6) Assuming complete expansion in the nozzles, what should be the

throat and exit areas of both nozzles? What should be the area of the
reversing chamber?
16. A reaction turbine passes a given weight of steam at its designed speed
under certain conditions of initial pressure, quality, and back pressure. If
the speed is reduced, what will be the effect on the quantity passed, and
why?
INDEX
Blading, impulse, 92
angle at entrance, 118
increase of, 119-121
Absolute velocit}^, 96 angle at exit, 121
Accumulator, heat, 15 decrease of, 104, 121, 123
Acoustic velocity, 64 increase of, 122, 228
Adiabatic expansion, of air, 62 breadth, 124
of steam, 49 carry-over, 104, 232
Admission, partial or full, 84, 86, edge thickness, effect of, 116-117
215 allowance for, 126
Advantages of turbines, 44 coefficient, 168
Alternating current generators, efficiency, conditions for maxi-
speed requirements of, 12 mum theoretical, 104
Angle of deviation, 123 variation with steam speed, 135
effect on velocity coefficient, 133 form, 115
Angles, blade (see Blading). friction, 99, 132
nozzle (see Nozzle). height, 117, 125-128
Arc, nozzle (see Nozzle). increase at exit, 125, 127
A. S. M. E. Steam Table Committee, maximum, 120
limitation,
53 minimum, 128
Available heat, 48, 50 passage width, 116
Axial clearance (see Clearance). pitch, 115
flow, 10 radius of curvature of face, 115
pitch (see Pitch). shrouding, 165
thrust, impulse turbine, 105 speed, 135, 138
reaction turbine, 157 speed ratio, 99, 135, 137
tapered, 121
B thrust, axial, 105
velocity diagram, 96-99
Balance piston, 7, 157 warped, 121
Baumann's curves for stage effi- work, 100-103
ciency, 136 Blading, purpose, 92
Bearing friction, 179 Blading, radial, on Ljungstrom tur-
Blade efficiency (see Blading). bine, 11
fastenings, 168 Blading, reaction, 138
height relation to ring diameter, angle at entrance, 151
156 at exit, 152
materials, 169 axial thrust, 157
wing, 151, 248 breadth, 152
Blading, Baumann multiple flow or carry-over, 142, 146, 250
multiple exhaust, 7, 158 clearance, 154
267
'
268 STEAM TURBINES

Blading, reaction, edge tliickness, Clearance, 7, 154, 162
effect of, 150 leakage through, 154
coefficient, 168 steam flow through, 164
efficiency, conditions for maxi- Coefficients, blade edge thickness,
mum, 143 168
as nozzle, 145 discharge, 72
variation with blade speed, 146 steam velocity in nozzle, 70, 72
end tightening, 163 curve of, 77
form, 140, 148 experimental determination of,
friction, 145 73
gaging, 149, 150, 247 steam velocity in blades, 99, 101,
height, 153 132
limitation, 153, 155 Combination turbine, pressure and
variation along the drum, 153, velocity stage, 6
246 velocity stage and reaction, 7
lashing, 166 Compounding (see Staging).
pitch, 150 Compound turbine, 9, 183
axial, of rows, 152 Condensers, vertical, 183
speed, 148 Condition curve, 79, 193
speed ratio, 143, 146 for different turbines, 194, 203
stage, definition of, 140 for 70 % and 80 % stage efficiency,
thinning at tip, 167 197
velocity diagram, 138 form of, 192, 200
warped, 155 relation to efficiency, 200
work, 141 Robinson's method, 195
Blading, velocity stage Conical casing and rotor, 117, 156
angles, 122, 130 Control, nozzle (see Governing).
breadths, 130 Convergent nozzle (see Nozzle).
efficiency, 112, 136 Corrosion, 46
stage, 137 60
Critical pressure,
height, 131 experiments on, 79
height ratio, 132 perfect gas, 61
pitch, 116 Critical velocity, of gas, 64
speed ratio, 111, 116, 136 of steam, 66, 67
velocity diagram and nomencla- explanation by Osborne Reynolds,
ture, 108-111 64
work, 112 Cross compound, 9
distribution, 112-113 Cross-over pipe, external, 8
Bleeder turbine, 14, 248 Cumulative heat, 190
Buckets, 2, 12 curve of, 225
By-pass governing (see Governing). Curtis turbine, Curtis stage, 6, 114
Cut-out governing (see Governing,
nozzle control).
Carbon packing, 4 Cycle, steam turbine, 51
Carry-over (see Blading).
Casing, conical, 117
D
Charts, heat, 50-54 Definition,steam turbine, 2
Classifications of turbines, 2 nozzle, 48
.
INDEX 269
Definition, steam turbine, blading, Efficiency, ratio, 188

92 rotational, 189
Design, data required, 219 stage, 137
nozzle, 19, 56
less in wet region, 200
thermal, 188
preliminary, final stages, 233
EUenwood charts, 53
pressure stage impulse turbine,
End point, 193
220
calculation of, 201
reaction turbine, 249 End thrust (see Axial thrust).
velocity stage impulse turbine, End tightening, 163
238 Enclosing wheel, effect of, 178
Deterioration of turbines, 46 Energy equation, 49
Deviation, angle of, 123, 133 Energy loss in blading, 134
Diagrams, heat, 50-54 Enthalpy, 50
velocity (see Blading). Entrance angle (see Blading,
Diaphragm, 5, 85 impulse, also Blading, reaction).
nozzles, 85 Erosion, 46
Disc friction, 178 Exhaust loss, 183
Discharge, nozzle, 55 turbine, 14
coefficient of, 72, 76 Exit angle (see Blading, impulse, also
Distribution of heat, 114 Blading, reaction).
in reaction turbine, 252 quality of steam at, 71
in ten-stage impulse turbine, 223 Expansion, an, 246
in two stage turbine, 240 path or curve (see Condition.
of reheat in impulse stage, 243 curve)
of work in velocity stage, 112 over and under, 60, 79
Divergence of nozzle walls, 82 experiments on, 79
Divided flow turbine, 9 losses due to, 81
DouV:)le flow turbine, 8 reversibility of, 49
Double reduction gear, 13 Expansions in turbine, number of,
Drum, 246 113, 252
Dummy pistons, 157, 172 Experiments on nozzles, 71, 80
leakage, 158, 175 Extraction turbines, 14
Economy, curve, 217 Factor, reheat, 190-191

variation with load, 208 Feed heating by bled steam, 14, 248
Efficiency, 186-191 Flow diagram of 3-stage turbine, 187
blading (see Blading, impulse, Flow, of perfect gas, 61, 64
also Blading, reaction). of steam, 55, 66
generator, 184 of steam and compared, 67
air
hydraulic, 189 velocity 97
of, 55,
internal, relative or engine, 188 variation with inlet pressure,

curves of constant, 197 206
leakage, 174 Friction in blading (see Blading,
mechanical, 189 impulse, also Blading, reaction).
nozzle, 70 in nozzles, effects of, 69
270 STEAM TURBINES
G Leakage losses (see also Clearance).
blade tips, 154, 170, 177
Gaging, 149, 150, 247 diaphragm packing, .174
Gearing for turbine drive, 10, 13 calculation for 10-stage turbine,
Generator, efficiency, 184 230
overloads, 214 dummy piston, 175, 185
speed requirements, 12 calculation of, 257
Governing, aim, 205 efficiency, 174
by-pass, 216 shaft packings, 172
combination, 218 Ljungstrom turbine, 10
general considerations, 205 Loads, most efficient, 205
nozzle control or nozzle cut-out, no-load, 209
209 overload, 214
on first stage, 212 generator, 214
throttling, 210 Losses, bearing friction, 179
effect on condition curve, 211 drum friction, 179
Governor, 4 electrical, 184-185
valve, 4, 205 exhaust, 183
Guide blades, 4, 130 leakage (see Leakage losses).
leaving, 182, 185

H mechanical, summary, 185
over- and under-expansion, 79
Heat charts or diagrams, 50-53
radiation, 182
content, 50
rotational, 177
distribution (see Distribution of
disc friction, 178-179
heat).
calculation for 10-stage tur-
Height ratio, 132
bine, 230
fanning, 178-179
steam pressure, 181
Impact method of measuring veloci- throttle drop, 181
ties, 72 variation with load, 207
Impulse, 3, 92, 141 water sealed gland, 180-181
experimental determination of, 73 Low pressure turbine, 5, 14
Impulse blading (see Blading,
impulse). M
Impulse turbine, 3
simple, calculation 105 of, Manufacturing system, turbine, 44,
Increased angles (see Blading,
87
imjmlse, also Blading, reaction). Materials of blading and nozzles,
Index for expansions, 63 87, 169
Intermediate or guide blades, 4, 130 Maximum blade height, 120, 155
blade speed, 138, 148
K
gas discharge through nozzle, 65
Keenan, Dr., steam chart, 53 Minimum blade height, 128, 153
Mixed pressure turbine, 13
L Moisture in steam, effect on condi-
tion hne, 200
Labyrinth packing (see Packing).
effect on stage efficiency, 200
Lead, nozzle, 132
.
INDEX 271
MoUier diagram, 52
Most efficient load, 205
Packing, carbon,4, 171-172
Motion, absolute and relative, 93,
diaphragm, 173-174
96
Multiple flow low pressure blading,
dummy or balancing piston, 173-
175
7, 158
high pressure gland, 172
N labyrinth, 170-173
losses (see Leakage losses, also
Net heat drop, 71 Losses)

Non-condensing turbine, 13 low pressure gland, 174
Nozzle, angle of divergence, 82 water seal, 170-171
angle of position, 83 used for high pressure, 181
increase of, 227 Parsons turbine, 6. 139 (see also
arc, 125 Blading, reaction).
block, 84, 87 Peripheral admission (see Admis-
calculation of ideal, 56 sion).
chamber, 4 Pitch (see Blading, impulse, also
control valve, 4 Blading, reaction, also Nozzle).
convergent, divergent, design, 68 axial, rows of reaction blades, 152
critical pressure (see Critical pres- Preliminary calculation of final
sure) . stages,233
definition and purpose, 48 Pressure, connections in turbine, 207
design, 68 critical (see Critical pressure).
effect of reheat on, 78 inlet, 206
diagrams, 58 throat, 59
diaphragms, 85 ratio, 62
efficiency, 70 staging, 5, 113
exit area, 78 variation in turbine stages, 206,
experiments, 71, 80 217
form, 58, 84 Propeller and turbine speeds, 13
friction, 69
lead, 132
Q
materials, 87 Quality of steam, 51
partitions, 85 at nozzle exit, 71
pitch, 117-118
plate, 4, 87
reheat, 70 Radial blading in Ljungstrom tur-
throat, 69 bine, 11
velocity coefficient, 70 Radial flow, 10
velocity at exit, 55, 71 Radiation loss, 182
Rankine cycle, 51
O Rateau turbine or stage, 5, 114
Rating of turbine, 214
Outages, 45
Ratio, blade height to diameter,
Over-expansion (see Expansion).
120-121
Overloads, 214
blade speed (see Blading, impulse,
generator, 214
also Blading, reaction.).
Over-speeding, 119
.
272 STEAM TURBINES

Ratio, efficiency, 188 Shock at blade entrance, 118-120
height ratio of blading, 132 Shrouding, 165
Reaction, 94, 138, 141 Simple impulse turbine, 3
blading (see Blading, reaction). design of, 105
method of measuring steam Single-double-flow turbine, 8
velocity, 73 Single- flow turbine, 8
one hundred per cent, 140 Single-pass turbine, 12
stage defined, 140 Single reduction gearing, 13
turbine, 6 Single stage turbine, 3
calculation for, 249 Speed, of rotation, 12
general considerations, 246 Speed ratio (see Blading, impulse,
pure, 144 also Blading, reaction).
work distribution, 247 Stage, 5
Reduction gearing, 13 efficiency, 137, 189
Re-entry turbine, 12, 123 calculation of, 221
Regenerative turbine, 14, 248 leakage, 137
Reheat, blading, 127, 243 Staging, 107-108
distribution of, 243 pressure, 113
factor, 190-191 velocity, 4, 108
effect of varying number of Steam formulae, 54
stages on, 194, 199 Steam Nozzle Research Committee,
effect of varying stage efficiency 73
on, 195 Steam rate with varying load, 208,
effect of varying superheat on, 217
198-199 Steam strainer, 4
efifect of varying initial pressure Steam Tables Committee, A. S. M.
on, 198 E., 53
nozzle, 70, 77 Steam turbine, definition, 1
effect on nozzle design, 78 advantages, 44

Reheating turbine, 15, 183 classification, 2
Re-superheating turbine, 15, 183 standardization, 1
Relative velocity (see Velocity). Stepping of reaction drum, 7, 57,

Repairing turbines, 146 247
Repeated flow turbine, 12 Supersaturation, 63, 89
Reversibility of expansion, 49 calculation of, 91
Reversing chamber, 12 limit, 89
Ring, blade, 115, 125 Superheated steam, expansion of, 63
Rotation, speed of, 12 Symbols, general, ix-x
losses (see Losses, rotational) heat and efficiency, 187
Rotor, conical, 156 velocity, 96, 109
stepped, 7, 153, 246 Synchronous speed, 118
hollow, 8
Rubs, 45
S Tandem compound, 9
Tangential flow, 12
Searching tube experiments, 80 Temperature changes, effects of, 46
Servo-motor, 205 Temperature-entropy diagram, 51
INDEX 273
Thermal efficiency, 188 Velocity, diagram, impulse stage, 96

potential, 50 reaction stage, 139
Thickness of blade edges (see Blade velocity stage, 4, 109
edge thickness). of flow, 98
Thinning at blade tips, 167 air,calculation for, 62
Throat, 59 (see also Nozzle). steam, calculation for, 55
Throttling control or governing, steam, experiments, 72
210 of whirl, 97
Throttle valve, 205 relative and absolute, 96
Thrust bearing, 158 stage, 108
Thrust on rotor (see Axial thrust). calculation for, 238
Tip leakage (see Leakage).
Total heat of steam, 50 W
U Water rate curve (see Steam rate
curve).
Undercooling (see Supersaturation] Water seal shaft gland, 180-181
Under-expansion (see Expansion). losses in, 180
Under-speeding, 119 Whirl, velocity of, 97
WlUans line, 208
Wilson line or limit, 89
Wing blades, 151
Valve, governor, 4, 205 Work, of simple impulse turbine,
nozzle control, 4 100-103
primary, secondary, tertiary, 213- reaction blading, 51
214 distribution in, 247
Vanes (see Blading). velocity stage blading, 112
Vector diagram, 96 distribution in, 113
Velocity, blade (see Blading).
coefficient (see Coefficient).
critical, perfect gas, 64
steam, 66 Zeuner's equation for adiabatic
equal to acoustic velocity, 64 expansion, 63
TJ735.C4 1928
<
a
LlJ
TJ735
C4 Church, Edwin Fayettet 1879-
1928 Steam turbines [by] Edwin F. Churchf
Jr. 1st ed. New York [etc.] McGraw-Hill
book cooapanyt inc* y 1928.
273 p* incl» illus* 24 cm»
18860
)
MENU 18 MAY 81 555149 NEDDbp 28-1292

Steamturbines Church

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Steamturbines Church

Uploaded by

Copyright:

Available Formats

Prof Alfred vi Ferrotti

McGRAW-HILL BOOK COMPANY, Inc.

PRINTED IN THE UNITED STATES OF AMERICA

THE MAPLE PRESS COMPANY, YORK, PA.

working knowledge of the ordinary steam cycles, the ability to

o< to be omitted for the sake of brevity.

^classified under chapter headings. It has been the author's

Manufacturing Company, the De Laval Steam Turbine Com-

II. Flow of Steam in Nozzles 48

IV. Losses 170

V. Efficiency, Reheat Factor, and Condition Line 186

VI. Output Variation and Governing 205

VII. Design 219

b = breadth of blading, inches.

h = height or length of blades, inches.

e = epsilon = efficiency (see Sec. 72).

K = kappa = ratio of specific heats = adiabatic exponent,

p = rho = ratio of blade speed to steam speed.

= theta = per cent of circumference occupied by active nozzles.

TYPES AND CHARACTERISTICS

Turbines have developed to their present forms along

2. Classifications of Steam Turbines.

^4. With respect to form of blade passage:

C. With respect to direction of steam flow relative to plane of rotation

TYPES AND CHARACTERISTICS 3

D. With respect to repetition of steam flow through bhides

E. With respect to rotational speed

F. With respect to conditions of use

These will now be considered in detail.

A. With Respect to Form of Blade Passages.

as applied to turbine blading will be considered later; for

and the diagrams compared, it will be noted that the steam

each being called a ''stage"; the resultant steam velocity

Figure 18 shows a view of a pressure-stage turbine during

the surface of a rotor in the form of a cyhnder stepped in

high-pressure drum of the straight reaction turbine (com-

blading in cases where it is desired to balance partially the

large area for flow through the two sets of low-pressure

blading. Figure 32 shows a view of a double-flow rotor,

which the steam passes in sequence through two or more

relation dictated by frequency requirements. When run-

to the axis of rotation among blades which are set radially.

near the shaft and flows substantially radially outward

are shown assembled and connected to the inner and outer

the steam commences to expand at a greater rate, requiring

cient, so that radial blades are necessary. These are

ble speed, as no generator can have than two poles. A

between the turbine and the driven machine, so that the

volume. This great increase in steam volume calls for

water for condensing purposes is unavailable, or where all

moderate sizes only (Figs. 8, 19, 21, 23, 26).

or turbine. The low-pressure steam enters and joins with

are utilized in manufacturing plants where steam at about

phragm opened and closed by a governor-controlled valve;

nor opens the valve wide and the turbine operates as a

several ways. First, when very large powers are required

from a single machine, it is sometimes advisable to divide

demand for more economical power-plant operation. High

^Text continued on page 44.

^A- ri^ ?^i2

Fig. 1. — Simple Fig. 2. — Velocity stage. Fig. 3. — Pressure stage,

Fig. 4. — Pressure and velocity stage. Fig. 5. — Reaction. (Parsons.)